tech

2025-02-19 · open the original ↗

• Deep learning hasn't helped tabular data because its features already exist. Hutter says tabular datasets are mostly small and unlike one another (omics blood values for Alzheimer's against transactions for fraud), lack the spatial regularity that convolutional networks exploit, and come with human-designed features; a data scientist's higher-level engineering, such as computing BMI from height and weight, is what adds value.
• TabPFN first beats XGBoost and random forests on their own information. He says the first step was an apples-to-apples test using only raw values and labels, not column headers; combining it with language models that read headers is in progress with "some initial results", and he expects it to be "so much better".
• TabPFN does tabular learning by in-context learning, like GPT. Hutter says the whole training set and test inputs go into the prompt and the model outputs the test labels, trained with cross-entropy, so each dataset is one training example; lacking trillions of quality real tables, the team generated all training data synthetically, which he believes makes theirs the first paper to reach state of the art on synthetic data alone.
• Sampling enough datasets from a prior makes a Transformer exactly Bayesian. He cites the team's 2022 paper "Transformers Can Do Bayesian Inference": a model fitted to many datasets drawn from a prior approximates the posterior predictive under that prior, and with an arbitrarily large Transformer and minimal loss it recovers exact results, such as true Bayesian linear regression or a Gaussian-process posterior.

2025-02-19 · open the original ↗

• "Squishy" human experiences can be measured and predict real outcomes. Hicks, who runs Pluralsight's Developer Success Lab, says psychology builds scales tested across thousands of people, likening its measurement error to atmospheric dust clouding starlight. Sense of belonging predicts persistence and grades, though she stresses rules have limits: belonging will not keep someone in a field that does not pay for groceries.
• Belonging on a team tracks with reported productivity. In her study of 1,282 developers, those on teams with a high sense of belonging and time to learn reported being more productive over the last month and rated their teams more effective. She says the lone-genius, 10x-developer stereotype holds grains of truth but, over-activated as a team norm, leads teams in the wrong direction.
• Code review anxiety competes with code for working memory. Hicks says anxious reviewers avoid or mentally check out of reviews, because worrying about colleagues uses the same cognitive resources as problem solving. Self-efficacy, the belief you will get past the friction, predicted productivity in her research, and she favors brief, repeated exercises such as her lab's workbook that build a "virtuous cycle" rather than a one-off intervention.
• Hackathons let people safely try on new roles. Pluralsight engineers asked her lab to measure their hackathon's value, and the lab found that people on supportive teams were more willing to try new skills afterward, whether or not product features resulted. She argues the temporary team lowers the stakes of experimenting with a new identity in front of a manager.
• AI skill threat is common, and team culture halves it. Surveying over 2,000 developers and 795 engineering managers, she found roughly 43% to 45% agreed they would not keep up or be judged fairly in an AI-coding future, and being a manager did not protect against it. Those on teams with belonging and learning cultures had about half the threat; she says cognitive offloading can be good or bad, the answer is "never going to be in the tool", and she worries tool builders are not bringing in learning scientists.

2025-02-18 · open the original ↗

• UniFi Network 9 replaces interface-based firewall rules with zones. The host lists the release's other additions as Site Magic SD-WAN, a local network API and UniFi CyberSecure, a $99-a-year threat-management service. Dave Barger, who presents the rest, says the old LAN In/LAN Out/LAN Local rule sets grew long and had to be read top to bottom to see the default action.
• The new zone matrix mirrors Crosstalk's own client firewall planning. Barger says Crosstalk consults clients on a source-versus-destination matrix coloured red (drop), green (accept) and yellow (complicated), and UniFi's zone matrix shows the same at a glance: "allow all", "block all" or a count of policies for each pair.
• The upgrade is opt-in and migrates existing rules automatically. After updating to 9.0 it must be enabled under Settings > Security; Barger's rules migrated in a second or two, though he assumes larger rule sets take longer. Networks are auto-sorted into Internal, External (WAN), Hotspot (guest) and DMZ zones, and custom zones can be created.
• Defaults allow internal-to-internal and block DMZ and hotspot traffic. External-to-internal holds three rules read top-down: allow return traffic requested from the LAN, block invalid traffic, then block all.
• Policies can now target a specific device, tracked by MAC. Sources can also be a network, IP or MAC, with ports and an allow, block or reject action. Barger warns iPhones and some Androids randomize MACs, so device targeting suits servers and NAS. Connection state is simplified to all states, return traffic (established/related) or custom.

2025-02-18 · open the original ↗

• Package suggestions from coding assistants are a supply-chain risk. Thomson gives three reasons: models can favour popular packages over good ones (he says the most popular Python OAuth package is no longer maintained and points users elsewhere), they know nothing of vulnerabilities found after their training date, and they can hallucinate packages. He cites security research in which a researcher registered a package a model had hallucinated, so the model then pointed users to a package the researcher owned. He says a malicious actor could do the same.
• CodeGate sits on the developer's machine, between the IDE and the model. Thomson's employer Stacklok started work on it at the end of 2024 as an open-source project. Its first feature checks the packages a model suggests against the data set behind Stacklok Insight, which tracks CVEs and maintenance across ecosystems such as npm and cargo, and shows a security warning in the IDE. He says most prompt gateways run on a remote server, while this one runs locally. Asked about revenue, he says that for now there is none: no services or freemium tier, just building the project and its community.
• Assistants upload context the developer does not control, including secrets. Thomson's example is a .env file holding API keys for a staging server or database. The assistant may send it as context, and .gitignore does not stop that. Even if the vendor is trusted, he does not want the data leaving the machine, passing through a corporate inspecting proxy or ending up in a log file. He also notes that some vendors promise not to train on context only on their paid plans. CodeGate replaces such values with tokens before sending and swaps them back on the way in. He says Stacklok is extending this from API keys and secrets to PII such as email addresses.
• Developers verify AI output less as they grow used to it. Thomson says that in talks with developers, senior ones "trust but verify", checking suggested packages with scoring tools or the registry, and some use Copilot in place of a search engine. Less experienced and "AI native" developers trust the output more. Giard adds that deadline pressure also cuts vetting, and he frames Copilot's "co" as a reminder that the developer is still the pilot. CodeGate is pitched as automating that manual vetting step.
• CodeGate workspaces route each project to its own model and prompts. Thomson's examples: Copilot for JavaScript and Ruby, a different model for COBOL, or ChatGPT for documentation if it handles English better. A workspace can also add a project-specific prompt, such as one for Next.js and React or one for Ruby.

2025-02-18 · open the original ↗

• TabPFN beats boosted trees by learning tabular prediction in context. Hutter says it feeds the whole training set and test inputs into a Transformer prompt and outputs test labels, using the same raw values XGBoost and random forests get and no column headers; adding language-model reading of headers is in progress with "some initial results".
• Training on synthetic datasets alone makes the Transformer approximately Bayesian. Hutter says the model learns from about 130 million generated structural causal models, the prior devised by first author Noah Hollmann, and approximates the Bayesian posterior over them in one forward pass; he believes it is the first state-of-the-art model trained only on synthetic data, and says it rules out test-set leakage, after reviewers rejected the first v1 submission as too good to be true.
• Version 2 turns a proof of concept into a general tool, for small data. Hutter says v1 handled only numerical classification on about 1,000 points and 100 features, and saw around 15 published uses; v2, published in Nature in January, handles regression, categorical, missing, outlier and imbalanced data, with text through the API only, and was evaluated up to 10,000 points and 500 features. He is confident of scaling to 100,000 or a million with combined approaches.
• Untrained on time series, TabPFN still leads time-series forecasting. Hutter says casting timestamps as features (time of day, day of week, day of month and sine and cosine encodings) lets TabPFN forecast any horizon in one pass; at recording it ranks first on the GIFT-Eval benchmark, ahead of Amazon's Chronos and other time-series foundation models.
• Hutter launched Prior Labs because academia can't ship for the masses. He says Auto-sklearn showed that a lab with a few research engineers builds only for researchers; Prior Labs, where he is co-founder and CEO, will build products while his academic co-affiliation pursues interpretability, failure modes and moonshots. The host says the TabPFN library had been downloaded almost a million times at recording.

2025-02-17 · open the original ↗

• Fotsch treats exercise as a productivity tool for desk careers. She says a fixed 30- or 45-minute workout block improves discipline, time management, energy, sleep and confidence at work, and that tech culture undervalues this as a help to being a better data engineer or analyst.
• Small daily movement adds up more than people expect. She recalls logging about 800 steps one day in her first data analyst job despite having worked out, and recommends stairs, walking Zoom calls and walking one-on-ones; someone managing one workout a week should aim for two plus more walks, not jump to six.
• Over-trainers should save all-out efforts for a few sessions. Responding to the host, who says copying two-hour, five-to-seven-day CrossFit regimens leaves him too tired to work, she suggests one or two all-out workouts a week. Otherwise, she says, scale intensity by how you feel and by readiness and sleep scores (taken with a grain of salt), progress gradually, and choose a programme that fits your own schedule rather than someone's on Instagram.
• Consistency, not perfection, is what produces fitness. Fotsch says fitness comes from many days of good choices, made about 80% of the time, rather than one hard week, and that goals can be as simple as feeling good and enjoying training; she saves her hardest efforts for Saturdays so weekday work doesn't suffer.

2025-02-16 · open the original ↗

• After retiring, Fotsch had to stop training like a Games competitor. She says that for months as a full-time analyst she still trained as if heading back to the CrossFit Games, with her identity tied to being strong at elite movements, until she asked whether muscle-ups and handstand push-ups served her current goals and began training for feeling fit "20 30 years from now".
• Short, minimal-equipment workouts got results without hours in the gym. Fotsch says she found she did not need her old intensity to reach her new goals, and that the dumbbell, bodyweight and hotel-room workouts she posted resonated with followers.
• PRVN GO drops competition skills for busy professionals and parents. Fotsch says PRVN's CEO, Nick Johnson, approached her soon after she joined CHG; unlike PRVN's competition-oriented CrossFit programming, the company of champion Tia-Clair Toomey and her husband Shane, PRVN GO omits high-skill movements in favour of conditioning, bodybuilding and core work. The host says the episode is not sponsored.
• A remote community supplies the accountability a class would. She says users who can't get to a class log the same workout and talk with others doing it elsewhere, which counters the excuses that build up after a hard week.

2025-02-15 · open the original ↗

• dbt turns raw warehouse data into stakeholder-ready models, mostly in SQL. Fotsch describes her work at Podium as adding fields and business logic to existing dbt models, later building models from scratch, using dbt's functions to consolidate repetitive code ("SQL on steroids"), and exposing the results in Tableau or Sigma.
• Documented definitions force agreement on what the business logic means. She says dbt docs make every field carry a definition agreed with stakeholders, which matters because something like "created date" or "active customer" may not mean what people assume, and which helps new engineers learn the data's meaning and granularity.
• Generated documentation and doc blocks remove the tedium. Fotsch says dbt now auto-drafts field definitions, which the host calls a good generative-AI use that removes the blank-page problem though output should be checked; doc blocks let a field defined in a staging model be reused throughout downstream lineage. The host contrasts the generic AI fitness tips he found useless on his Whoop.
• Sigma's spreadsheet feel won adoption from product managers. She says Sigma sits directly on Snowflake, removing an ingestion step, and its Excel-like look served both analysts and Excel-familiar PMs, which helps get everyone onto the same numbers.

2025-02-14 · open the original ↗

• Plan with compounding percentages and feedback loops, never linear increments. In a clip from episode 855, futurist Azeem Azhar says nothing grows linearly except birthdays: growth follows S-curves with an exponential phase. He cites the Intel–Windows upgrade loop that drove chip progress, builds dynamic percentages and feedback loops into his own forecasts, and says a linear forecast of S3 storage spend makes "incredibly extreme assumptions", even if the realistic one alarms the CFO.
• Hybrid quantum computing targets scheduling and logistics that classical heuristics only approximate. In episode 851, Florian Neukart describes a Volkswagen production-testing problem with 250-odd test classes, each with subtests, to fit to available staff skills; he also describes fleet routing with delivery windows and empty miles. He concedes quantum results are not provably optimal either, so the classical and quantum answers are compared and the better one used.
• Choosing a foundation model means weighing a long criteria checklist. In episode 853, Kirill Eremenko lists twelve: cost, modality, customization, inference options, latency, architecture, performance benchmarks, language, size, scalability, compliance and licensing, and environmental impact; Hadelin de Ponteves says in practice his choices, such as an Amazon Titan model and a Llama model, came down to which matched his fine-tuning dataset's format.
• Layered, reference-free metrics evaluate agents better than single accuracy scores. In episode 857, Coval's Brooke Hopkins says getting the most-used case wrong matters more than a 99% average, so metrics are combined: a short conversation is bad only if the goal also failed. Coval checks JSON-described workflow steps and function-call arguments, inferring correctness from context rather than golden datasets, and flags agents that take longer than humans or repeat themselves.

2025-02-13 · open the original ↗

• Analytics engineering is the data-modeling layer between ingestion and analysis. Fotsch says that at Podium, data engineering pulled data in and maintained pipelines, analytics engineering modelled the raw data in dbt for stakeholder use, and analysts built Tableau visualizations, working with the modellers on the questions being answered.
• Whether the role exists depends on the company. She says neither CHG, her current employer, nor Homie had a separate analytics engineering team, though the work still happens; at Podium, analysts gradually took on analytics engineering work as the team grew, which she credits with leveling up her technical skills.
• Knowing what tools exist matters more than having used them all. Fotsch says learning from what fails, and knowing where to look, is the core of the job; on her first day she wanted to take ten courses at once, and her sister-in-law, a data analyst, told her it was a marathon, not a sprint. Context on the wider pipeline, she says, helps with stakeholders and with knowing whom to ask.

2025-02-12 · open the original ↗

• Formal training builds the confidence to explain your value in interviews. Fotsch says a degree such as her Arizona State programme, plus Coursera courses she still uses, helps both on the job and in interviews; for analyst roles she prioritized SQL above all, some Python, and the popular BI tools Tableau and Power BI, whose skills transfer.
• A referral from someone who knows you conveys what a resume can't. She got her Homie interview through a friend from the fitness world working on the product side, who connected her to the data manager and could vouch for her work ethic and teamwork; she advises listing target companies and industries and finding anyone you know there, not only people in data.
• In-person meetups build a community worth having even when employed. Fotsch says Utah has many data meetups and the contacts she makes there are people she can later ask how they solved a problem; the host adds that newcomers with little knowledge are welcome at data analyst, data science and AI meetups listed on Meetup.com.

2025-02-11 · open the original ↗

• The module automates the click-ops of setting up cloud operations. Tinderholt, who works in Azure Core Engineering, describes the GitHub AT-AT as an opinionated Terraform module library: running terraform apply locally, with access to GitHub, Azure and Entra ID, creates a GitHub repo holding Terraform HCL code and GitHub Actions YAML, and sets up Entra ID federated identity credentials so those workflows can reach Azure. He calls it "using Terraform to Terraform": it uses the GitHub, Entra ID and Azure providers to drive those control planes rather than to provision VMs.
• One foundation supports radically different release motions. Besides the basic "starter app", he has extensions for Azure Functions, a Jekyll static website and a VM workload that builds a Packer image and deploys it to an Azure VM. He runs his own blog, marktinderholt.com, on the Jekyll module (a pull request triggers a pipeline that publishes to Azure Storage) and stood up a Minecraft server with the VM module.
• The generated workflows build in a standard Terraform GitOps process. He says the GitHub Actions let you run a plan manually to check drift, or follow the full flow: open a pull request, review the plan it shows, and apply on merge, using pre-compiled plans. A team that already has a repo can pull it in with a Terraform data source and pass it to the module instead of having it created.
• Isolation is built in but optional, "choose your own adventure". Because Terraform state, kept in an Azure storage account, can hold sensitive values, he recommends isolating production and non-production backends. Multi-subscription support also brokers RBAC so each environment's identity holds only the privileges it needs there; someone just kicking the tires can skip both.
• It suits teams adopting Git flow, not lone wolves. He says little Terraform knowledge is needed, but users must know Git and work through feature branches and pull requests; someone working alone is probably better off running Terraform locally. He is the sole contributor but open to pull requests, publishes the module on registry.terraform.io, and says a sibling Azure DevOps AT-AT exists but has not been published in a while, though users have asked him to bring the GitHub version's features back to it.

2025-02-11 · open the original ↗

• Fotsch reached the CrossFit Games after years of near misses. The host reads her record: 491st worldwide in the 2014 Open, 113th in 2015 (when she won workout 15.5 outright), 80th in 2016 but seventh at Regionals with five qualifying, fourth at the 2017 Games on a team after an 80% rotator-cuff tear, seventh again at 2018 Regionals, then 33rd in the 2019 Open and 27th at the Games; before that she won two NCAA titles with Cal and twice broke an American relay record.
• A flawed Games appearance pushed her to plan life after sport. Fotsch says the 2019 Games' expanded field of national champions, upwards of 100 per division against the usual 40, was not the Games athletes felt they deserved, and taught her she was not entitled to any outcome; elite athletes, she argues, avoid thinking about what comes next.
• She was already a data nerd from tracking her own training. Fotsch says weighing food, timing meals and sleep, and overlaying wearable data on workouts so her coach could balance load against recovery drew her to data, after her sister-in-law, a Netflix analyst, described the field; her kinesiology master's, she says, overturned a belief that she wasn't smart enough for technical work.
• An ACL tear ended her bobsled bid and hastened her data career. She made the US national bobsled team as a brakeman in 2021, aiming for 2022, but tore her ACL weightlifting months before qualification, and began job applications from bed after surgery while finishing an Arizona State business data analytics master's; a friend's referral got her an analyst role at Homie.
• Her path ran from analyst to analytics engineer to data platform manager. At Podium she moved from Tableau analysis into dbt modelling; at CHG Healthcare, a medical staffing firm, she is senior technical manager of marketing technologies within the data platform team, centralizing scattered marketing data, delegating more and building less. She credits CrossFit's "figure it out" habit for her comfort learning on the job.

2025-02-10 · open the original ↗

• Like a compiler, BAML handles the prompts nobody will hand-tune. Gupta says a decade of performance engineering taught him that compilers beat him on time per dollar, so a team should hand-optimize only its one or two most critical prompts and let tooling handle the other 90%.
• Schema-aligned parsing tolerates the model's malformed output instead of forcing strict JSON. Gupta says his team spent about eight months on an algorithm that coerces a response into the expected data model, fixing missing quotes and commas and strings given where numbers were expected, in under five milliseconds; he says customers save 20% to 30% of output tokens.
• The parser brings function calling to models released without it. Gupta says it needs no model modification, so BAML users were tool-calling with DeepSeek R1 and OpenAI's o1 models at recording, and a customer building a Cursor-like tool for Xcode had the LLM emit code in triple quotes, avoiding JSON escaping, which BAML turned into properly escaped strings.
• Semantic streaming lets chat products show a generative UI, not just text. He describes type-safe partial data driving interfaces such as recipe cards with a progress spinner per section, or live tickers and graphs for stock analysis, and argues a chat startup must ask why its users wouldn't just go to ChatGPT, with better UX as the answer.

2025-02-09 · open the original ↗

• Good RAG depends on data-specific retrieval no general tool can supply. Gupta says his company started on RAG pipelines as one of its 13 pivots but moved away because pulling the relevant documents depends on each dataset, so there is no general RAG solution, just as there is no universal web accordion component.
• JSON pasted into prompts is filler the model must filter out. He calls them the prompt's "likes and ums": quotation marks, escape characters and colons make context parseable for code but harder for the model to read, and that the placement, orientation, size and surrounding text of images also matter; BAML renders the actual prompt live in VS Code so developers can spot what looks ugly, on the reasoning that models are trained on human data.
• Treating prompts as typed functions curbs the contradictions of long English instructions. Gupta says every BAML prompt is a function with type-safe parameters and a declared return type, such as an enum of categories, with each category's rule attached to that category rather than one growing essay; this, he argues, yields shorter prompts written like code.

2025-02-08 · open the original ↗

• Gupta predicts prompt engineering will mature as front-end development did. Responding to a host premise, from his researcher, that it is a "joke discipline", Gupta says he once objected to the term "AI engineer" being used for prompt writers, but now sees prompting as being where web development was with plain HTML, no CSS and no interactivity, before it became a complex craft; he expects it to become a well-studied discipline over time.
• What prompting needs is reliability in transforming data. Gupta frames a prompt as a way to turn supplied data into the data you want, such as classifying a transcript as a doctor–patient conversation or general wellbeing, and says using a model involves more nuance than people credit.

2025-02-08 · open the original ↗

• DeepSeek run locally showed no traffic to China in the host's capture. Following David Bombal, who used Wireshark to show the DeepSeek smartphone app sends data to China, the host ran DeepSeek under Ollama and Open WebUI with Wi-Fi disabled and other apps closed, filtering for outbound TCP. A query produced almost no activity, and the only endpoints were in the United States.
• The local model answered with the network fully disconnected. With the Ethernet adapter disabled it still reasoned through and correctly answered an arithmetic question (12 apples).
• Local LLMs trade privacy for less processing power. The host concludes local models do not phone home to China or anywhere else, but lack the compute of OpenAI or DeepSeek; he still rates local running solid for general queries and document summarizing.

2025-02-07 · open the original ↗

• DeepSeek R1 matches the top models on human-rated leaderboards. The host reports that at the time of writing R1 was statistically tied for first on the overall LM Arena leaderboard, within a 95% confidence interval, alongside GPT-4o and Google's Gemini 2.0 Flash, and that its release coincided with Nvidia's share price falling 17% and the Nasdaq several percent in a day.
• R1's real shock is its tiny training cost. In the host's rough approximation, a DeepSeek V3 or R1 training run costs on the order of millions of dollars against hundreds of millions for o1, Gemini or Claude 3.5 Sonnet; he credits a conceptual breakthrough combining mixture-of-experts with new efficiencies such as DualPipe, a GPU communications scheduler, across a couple of thousand GPUs, about 1% of what Meta and xAI say they buy in a year.
• The host argues US chip sanctions on China have proven ineffective. He says a Chinese firm approached US capability with about 1% of the chips, at about 1% of the cost, using less capable Nvidia chips, and that DeepSeek released V3 and R1 source code and weights under the MIT license, more open than Meta's Llama, which shares weights but not code and restricts companies with over 700 million active users.
• Avoid DeepSeek's app, but run its models privately. The host warns that, per its privacy policy, anything entered in DeepSeek's iOS app, number one on the App Store at recording, is stored on servers in China; he suggests Ollama for private use.
• Efficient training may strand the giant capital raises, the host argues. In his view, recent $6 billion raises by OpenAI, xAI and Anthropic and the $500 billion Stargate project probably assumed models would keep growing by orders of magnitude; he welcomes cheaper training as better for the environment and for building applications.

2025-02-06 · open the original ↗

• Tailscale is free for personal use and built into GL.iNet routers. It is a WireGuard-based VPN and SD-WAN tool, free for up to 100 devices and three users, where each device in the tailnet gets its own private IP. On GL.iNet routers it is labelled beta, so older firmware may need upgrading; the host demos his favourite travel router, the Beryl AX.
• Binding the router takes a toggle and a link. Enable Tailscale under Applications, then follow the device-bind link to connect it to your tailnet. The Tailscale admin flagged the router's client as outdated; the host found no way to update it but saw no problem, and expects a future GL.iNet release to update it.
• Reaching devices behind the router needs two extra settings. Tailscale alone reached only the router's admin page. Enabling "Allow Remote Access LAN" on the GL.iNet and approving its advertised 192.168.8.0/24 subnet under Edit Route Settings in the Tailscale admin let him reach an OpenSpeedTest server at 192.168.8.150:3000 from a separate network.
• Remote Desktop into a home PC needs Windows Pro and a fix. Windows 10 or 11 Home cannot accept RDP; enable Remote Desktop and allow RDP through Windows Firewall. Login with a Microsoft account failed until he ran a "runas" command with his MicrosoftAccount email and winver, from an older article he links.
• The host says the whole setup is quick, even for first-timers. He puts it at no more than 10 minutes start to finish and says he was impressed; a written blog post of the steps is linked in the description.

2025-02-06 · open the original ↗

• Y Combinator partners doubted the Slack-competitor idea that got Gupta admitted. Gupta says he and his co-founder, UT Austin graduates of 2015, tried edtech and the creator economy before applying to Y Combinator with a warmer alternative to Slack; on his fourth application, and third interview, they got in, though the partners told them the idea would not work.
• Fielding batchmates' machine learning questions revealed the real market. He says they joined the Winter 2023 batch as ChatGPT came out, and as more seasoned engineers, one with nine years in distributed systems and the other nine or ten in algorithms and assembly, found themselves answering mostly ML questions; they also concluded they shouldn't be building UIs, and the company went through 13 pivots in all.
• An SDK can't express what's needed; a new cross-language syntax can. Gupta says Python or TypeScript SDKs such as Vercel's AI SDK and Pydantic AI are limited by their host language, and since every software shop, whether C++, Java or Go, will use AI, the answer is a new syntax like JSON, YAML or TOML before it, which became BAML.

2025-02-05 · open the original ↗

• Tails boots any PC from USB and leaves no trace. Building it needs a USB drive of at least 8 GB (USB 3 recommended for speed): download the image from tails.net, verify it, and write it with balenaEtcher, choosing the target carefully because it is wiped. The host PC may need its BIOS boot order changed to boot from USB.
• Optional persistent storage is encrypted with a passphrase. The first-run wizard offers it, and you choose which data persists between boots, such as documents, bookmarks and Wi-Fi passwords.
• Tor masks location and activity but is not bulletproof. Tails offers connecting through Tor and bundles the Tor Browser, Pidgin for secure messaging, and network-monitoring and encryption tools; rebooting and removing the stick ends the session.

2025-02-05 · open the original ↗

• Prep time scales with the stakes and how new the material is. Mark says after thousands of talks he preps little because his content is incremental, yet their joint prepared demos take 10 to 15 hours. Hanselman lists six talks in one week prepped from 10 minutes to a couple of hours, against a 45-minute AI Tour slot in Johannesburg before 5,000 people with backup machines, and a TEDx Portland talk of 15 to 18 minutes that is getting in his head.
• Design the talk for one defined target listener. Mark says to decide who the audience is and what they know, aiming at that person and briefly bringing others up with "for those of you not familiar"; not knowing the audience is when he has been flummoxed. Hanselman says to start from what people should leave with, such as something they can implement that day, and to remember the room wants you to win; both pick a friendly face on each side to play to.
• Time-box sections and keep the plan on paper. Mark divides his minutes across sections, rebalances them, writes clock times to check against, and memorizes slide order because unrehearsed transitions look amateur. Hanselman splits 45 minutes into 5-minute intro and outro bumpers plus seven 5-minute sections in three acts, keeps notebooks from every talk, and advises beginners to keep two or three talks ready, since in his experience one or two conferences in ten lose a speaker.
• Pre-stage nerves are physiology to manage, not a warning. Mark uses a power pose and biofeedback, watching his heart rate and breathing to get it under 100 before going on. Hanselman says his heart rate runs 105 to 110 before a talk and he reframes it as his body sending blood to his brain to help him succeed.
• Keep forward momentum instead of second-guessing words. Mark primes himself to speak slowly and accept the words he picks, likening stumbling to watching your feet while running downhill; one host says he has replaced "um" with silence over roughly the last 10 years, which they agree audiences do not notice. Hanselman credits Billy Hollis with staging tricks such as assigning client and server sides of the stage, and recommends Toastmasters as proof that speaking is learned technique.

2025-02-05 · open the original ↗

• BAML aims to do for prompts what React did for HTML strings. Gupta, CEO and co-founder of Boundary, says that as a Meta intern he saw how concatenated HTML strings let one unclosed div break the news feed, and that React's new syntax combining HTML and JavaScript gave instant error squiggles, state management and hot reloading; BAML, "basically a made-up language", applies that idea to LLM calls.
• Ever more prompts in codebases demand tooling that catches errors. Gupta predicts each year's codebase will hold more prompts than the last, and that models of all kinds will get easier for non-experts to use, which calls for developer tooling for consumers of models rather than their producers.
• Prompts need TypeScript-style type safety and static analysis. He argues that however good models get, their use is still code that must fit a codebase, and that BAML's algorithmic work reframes prompting the way TypeScript's type safety improved JavaScript.

2025-02-04 · open the original ↗

• Prompts need their own language, as HTML strings needed React. Gupta, CEO and co-founder of Boundary, says codebases will hold more prompts every year and that SDKs in Python or TypeScript are limited by their host language; BAML is a new cross-language syntax, like JSON, YAML or TOML before it, that treats each prompt as a typed function with typed parameters and a declared return type, such as an enum whose categories carry their own descriptions.
• A hot-reload loop multiplies how many prompts a developer can test. Gupta says that without BAML he sees people try about five prompts in 20 minutes, while BAML's in-editor tests, modelled on Rust's run-test buttons, take about 5 seconds each, allowing 240 in 20 minutes; it also renders the actual prompt live so developers can spot formatting noise such as JSON blobs and badly placed images.
• Schema-aligned parsing repairs malformed model output and saves tokens. Gupta says about eight months of work produced an algorithm that coerces responses into the expected schema in under five milliseconds, which he says saves customers 20% to 30% of output tokens and gives function calling to models released without it, such as DeepSeek R1 and OpenAI's o1, at recording.
• Next, BAML adds conditionals and versioned data schemas. Gupta says users want control flow, for example routing free users to GPT-4o mini and paid users to o1 while reproducing each in the test loop; a protobuf-like versioned schema layer is meant to stop rigid data pipelines from breaking when an enum grows from five categories to 50.
• Boundary hires on references, not technical interviews. Gupta says the company has never posted a job: candidates email "why I'm awesome" with three achievements, then have a short call, after which Boundary interviews three references in depth on the technical work. Those offered a job can spend a week in the office, five days a week in person, and Gupta offers to help them interview elsewhere, hoping they still choose Boundary.

2025-02-03 · open the original ↗

• Coval started with voice agents because they are constrained and booming. Hopkins says one person talking to another is a narrower medium than self-driving or web agents, so better metrics and workflows can be built for it; voice agents were also "exploding" at least six months before recording, and they extend call centers and phone trees companies already run.
• Voice gives any two businesses a universal natural-language API. She argues a company's agent can expose what it knows, such as claims data, appointment availability or opening hours, to any caller, whether that is a person on a traditional phone without an API or another agent, without integration maintenance.
• Hopkins predicts voice is the first big platform since mobile. Citing a Steve Jobs talk on how early TV filmed plays and early mobile shrank websites before native forms emerged, she expects every company to be expected to have a voice or chat agent, and new experiences that mix speech with web forms, for example for delivery drivers, police officers or truckers away from their computers.

2025-02-02 · open the original ↗

• Even near-human agents will need monitoring at scale. Hopkins says Coval's vision is to manage and understand how agents behave at scale even once they execute tasks exceptionally well, just as call centers and large companies still review human performance; understanding agents is what lets society reason about, legislate and deploy them for the right uses.
• Hopkins expects human creativity to build on agents, not yield to them. She predicts that 50 years from now will be as different as 100 years ago is from today, but that change is neither as dramatic nor as linear as it seems; if agents take over mundane work like spreadsheets and email, she argues, people will become more creative and connected, as they did with computers.

2025-02-01 · open the original ↗

• Agent quality is a spectrum, so Coval layers many metrics. Hopkins says no single score separates good from failing conversations; a suite of metrics shows trends and allows trade-offs, because getting the most-used case wrong, such as booking an appointment, matters more than a 99% average, unlike traditional machine learning.
• Combining metrics separates real failures from false alarms. She says a short conversation isn't bad on its own, but one that is short, missed its goal and skipped required steps is; filtering by metrics directs scarce human review to the most interesting examples.
• Workflow and function-call checks catch technically completed but wrong tasks. Coval checks conversations against JSON-described workflows, for example an appointment booked without asking for the phone number, infers the workflows agents actually follow in production, and verifies that tool calls were made with the right arguments.
• Reference-free evaluation suits non-deterministic agents better than golden datasets. Hopkins says reference-based metrics need curated expected outputs that must be maintained as prompts, models and infrastructure change, while reference-free ones infer the correct answer from context; borrowing from self-driving, Coval also compares agent and human task times, where much longer times plus repetition suggest an agent going in circles.

2025-01-31 · open the original ↗

• The podcast is hiring an account executive to sell sponsorships. The host says the role is open only to applicants based in Canada, ideally near the Eastern Time Zone, with at least 5 years of full-time sales experience, and owns revenue targets and sponsor strategy, with a technical AI or data science background a plus.
• The host says the show is far below its revenue potential. He states that podcasts with comparable listenership earn about 10x its revenue, after more than eight years, over 850 episodes at 104 a year, and three years of sponsorship. He suggests the high-production TV ads he began hosting in the past year for Nvidia, Dell and AT&T could become a sponsor offering.

2025-01-30 · open the original ↗

• Children should be creators of technology, not only consumers. Answering Hanselman's story of a family who found a 3D printer too expensive while sitting in front of a PlayStation 5 and its games, Berry says gaming is fine but should come with building: Minecraft worlds, Roblox, and the maker spaces and 3D printers now in public libraries.
• Art belongs in STEM, and may be the hook for girls. Berry argues engineers need history, ethics and a sense of how people meet their designs (the uncanny valley in robotics), and that ugly buildings and bridges don't get funded. She says sewing, cross-stitching or dolls do not disqualify anyone, and objects to parents at her robot exhibits going to fetch a girl's brother.
• High school homework builds the executive function engineering needs. She reports that since the pandemic, incoming engineering majors cannot self-organize or plan long-term, and wants homework for reflection whether or not it is graded. Because students are used to scrolling, she teaches her 50-minute lectures (phones banned) through active learning, not "chalk talks".
• AI is a tool to check, not a crutch. Unlike English professors who oppose ChatGPT on assignments, she wants engineers judging whether an answer is correct and efficient; she has seen ChatGPT give wrong answers to her exam questions. She compares it to a calculator: in AC circuits a student should be able to tell whether 350 volts makes sense.
• Students learn debugging by struggling through it themselves. She will not point out a forgotten power-supply switch or missing semicolon, because a student who spends five hours finding it never forgets; her own 17-hour session ended in a carriage return. She frames the skill as breaking a system into subsystems, and teaches multidisciplinary robotics because no design avoids another discipline; the broken-toaster exercise ("are the neighbours' lights on?") is Hanselman's illustration.

2025-01-30 · open the original ↗

• A ChatGPT-style local AI needs no powerful GPU. The host builds it from three free pieces (Ollama to run the models, Docker Desktop, and Open WebUI as the chat interface) on a Minisforum MS-A1 with an AMD Ryzen 7 8700G, Radeon 780M graphics and 32 GB of RAM, about $700 new; because the models run locally, he says, nothing is sent to corporations or governments.
• The WSL2 step tends to fail once before succeeding. Enabling the "Virtual Machine Platform" Windows feature (Windows 11, or Windows 10 version 1903 build 18362 or later) failed on his first attempt and went through on the second; Ollama itself installs with no GUI (the setup is about 750 MB), and Docker Desktop prompts a WSL2 update after its reboot.
• Pick the Open WebUI run command that matches your GPU. The default command from the Open WebUI quick start suits a PC without an Nvidia card; with one, the Nvidia-GPU variant offloads the model work and runs much faster. The interface comes up at localhost:3000, the first login creates the admin account, and the host stops the container when he wants the GPU back.
• Start with small models and step up until speed suffers. Llama 3.2 at its default 3 billion parameters ran well on the MS-A1 (the 1B version is pulled as "llama3.2:1b"); Llama 3.3 at 70 billion was "almost unusably slow" even on his Threadripper 3960X desktop, though DeepSeek R1 70B ran fine there, so he used DeepSeek R1 7B on the mini PC.
• DeepSeek's censorship sits mostly in its app, not the model. He reports the claims that DeepSeek R1 was trained for under $6 million, knocked about half a trillion dollars off Nvidia in a day, and (per OpenAI) was trained using ChatGPT. The local model had refused to discuss the 1989 Tiananmen Square massacre while he researched, but answered on camera; Ollama also lists uncensored models such as Llama 2 Uncensored.

2025-01-30 · open the original ↗

• Agent errors compound across pathways, so evaluation must be multi-step. The host poses a triage agent calling the wrong tool early; the Coval guest says multi-step evaluation differs from single LLM calls because non-determinism multiplies across every possible path, though agents sometimes recover on their own.
• Self-healing and redundancy give agents graceful failure. She describes a cheap, low-latency model answering while a slower "overthinker" reviews the whole conversation and prompts it back on track, plus fallbacks borrowed from aerospace and self-driving. She says the majority of Coval's customers can already hand off to a human when a task is too complex.
• Reliable agents can be built from unreliable parts, as cloud was. She points to servers being unreliable yet cloud systems reaching six nines through redundancy and fallbacks. The host goes further, calling agents handling complex tasks "inevitable"; she adds that each reliable enterprise deployment raises the tide for all, as with self-driving.

2025-01-29 · open the original ↗

• The workflow turns manual review into automated metrics. The Coval guest describes building an MVP, iterating on the prompt without voice, then running test sets of scenarios (book an appointment, issue a refund, complain about an airline experience) as 100 simulated conversations at once. Failures get reviewed by hand, what the developer looks for becomes a metric, fixes such as less eager interruption get re-simulated, and the loop then moves to production monitoring.
• No voice-agent platform fits every use case. She declines to name preferred providers, saying Coval aims to stay framework-agnostic so customers can switch as prices and requirements change. The axes she gives are low-code versus configurable (open-source orchestrators offer more control over function calling and RAG) and the weight placed on instruction following, function calling, workflow following, naturalness and creativity.

2025-01-28 · open the original ↗

• Microsoft Learn is now the umbrella for all Microsoft developer content. Ashcraft, who writes Windows developer docs for Microsoft Learn, says it grew out of MSDN and TechNet via Microsoft Docs, and now covers documentation, training modules and learning plans, certification prep and practice tests, and a Q&A site that Microsoft support staff also monitor. He describes a new "adaptive learning" feature in which AI builds a learning plan of modules and docs from what a user says they want to learn.
• The docs are written and shipped exactly like code. Ashcraft says content is mostly Markdown with some YAML, kept in GitHub repos (some API repos in Azure DevOps); once new APIs land in a Windows SDK build, a process generates empty Markdown stubs that writers fill in from the product team's API spec, then a pull request goes to product-team or peer reviewers, with GitHub Copilot now also available as a reviewer for typos. The site is built with the open-source DocFX plus Markdown extensions, and writers use VS Code with the public Learn Authoring Pack for preview and linting.
• Stale pages are archived out of search, and links are policed. Ashcraft says outdated docs first move to archive repos where they stay reachable by URL but drop out of Google and Bing indexes. A weekly Power BI report emails him the broken links in his area, and monthly "freshness" batches take top-viewed articles not updated in a while, 10 or 15 a week, to re-check.
• Readers can fix pages themselves or send feedback. Ashcraft says each page's pencil icon opens the GitHub source so a reader can submit a PR entirely in the browser, routed to the page owner; the thumbs-up/down button takes anonymous feedback without a GitHub login; and open-source products such as .NET keep GitHub issues open, which he prefers because it allows back and forth. His team triages issues, feedback and PRs weekly, prioritizing by how wrong the content is and how complex the fix.
• EU regulation is partly behind better Windows API documentation. After Giard says Microsoft docs were mocked 20 years ago and have improved over the last 10, Ashcraft says part of the improvement in the last couple of years was forced by EU rules requiring a certain level of documentation for all Windows APIs, with possible fines, which makes API coverage a high priority for his team.

2025-01-28 · open the original ↗

• Self-driving simulation methods transfer to testing voice and chat agents. Hopkins, who led evaluation job infrastructure at Waymo before founding Coval, argues that agents, like self-driving cars, must respond to a changing environment, so they need many simulated pathways rather than manual back-and-forth testing, which she says is what most companies do today. Coval simulates scenarios (e.g. "book an appointment for tomorrow") 10 or 100 times, and re-simulates from production transcripts the way Waymo replays road logs.
• Test coverage is a cost-and-latency trade against signal. Hopkins says Coval works with customers to decide what runs on every PR, every six hours or nightly, and before big releases, and to check whether test sets resemble the distribution seen in production.
• Layer reference-free metrics rather than chase one score. She describes workflow-adherence checks against a JSON workflow, function-call checks inferred from the conversation rather than a golden dataset, unanswered-question detection, and agent-versus-human task time. A short call is not bad on its own; a short call with an unmet goal is. Getting the single most common case wrong matters more than a 99% average.
• Reliability can be built from unreliable parts, as cloud infrastructure was. Hopkins cites redundancy, fallbacks, a slow "overthinker" model monitoring a fast low-latency one, and human handoff, which she says most of Coval's customers already have. She judges self-improving agents and automated prompt optimization still too early; visibility beats the time saved, for now.
• Voice becomes a universal natural-language API between businesses. She says Coval started with voice because it is more constrained and was exploding faster than other agents about six months before recording, and predicts voice is the first platform since mobile every company will be expected to have.

2025-01-27 · open the original ↗

• There is no solid ground for workers facing AI. The host premises that remote keyboard-and-microphone jobs, including much software and data science work, are easiest for AI to replace. Azhar tells listeners to feel apprehensive, and to spend two or three days throwing their hardest problems at LLMs until they are shaken, then "come out the other side".
• Jobs already turn over fast, and new ones keep appearing. He cites that 53% of Gen Z want to be influencers, 14 years after his own influencer-ranking startup, and economist David Autor's finding from about 80 years of US data that 60% of today's jobs did not exist in 1940. He notes the tension between middle-class desk-job growth and lower-paid service jobs over 30 to 40 years.
• Work is becoming "boss-like", and skilling up is the only preparation. Azhar says he and his team now set goals for teams of one or more AI systems, where domain knowledge and judgment help. He expects most jobs to change, some to vanish and new ones to appear, but says nobody knows how fast, and that even in 20 years the question may be open.

2025-01-26 · open the original ↗

• Energy is shifting from a commodity to a technology that cheapens. Azhar argues AI's energy demand will catalyse discipline in building the system. He says coal, oil and gas have not got cheaper in 200 years, while solar panel costs fell by, he thinks, 99% over 20 years, with batteries following the same exponential pattern and intermittency "a completely solvable design problem".
• The transition will mirror telecoms giving way to the internet. He likens the fossil system to the old phone network, reliable but needing about a billion dollars to enter, and forecasts a decentralised, cheaper system led by solar plus batteries, with wind, nuclear and maybe fusion. He puts the energy transformation at a couple of decades and cheap carbon capture and desalination at decades.
• The pessimistic case is not grounded in evidence, he argues. Azhar says peer-reviewed papers in Science and Nature show solar needs less material than fossil fuels, that lithium is plentiful but under-extracted, and that an Oxford study from about two years earlier found a fully renewable system cheaper to run, saving "trillions" the faster it is built. He likens the objections to a UK mobile CEO's 1999 vow never to allow online bill payment.

2025-01-25 · open the original ↗

• When quantum computing becomes useful is genuinely uncertain. Azhar says he does not know, and that quantum builders are scientists who see the world "as it could be", so their time frames differ from product builders'. He expects amazing things once machines have tens of thousands to millions of logical coherent qubits, and notes Google announced its random-number benchmark breakthrough twice.
• Transformers on GPUs may get there before quantum machine learning. He argues transformer architectures are "eating" protein, materials and robotics work, so the overlap with quantum machine learning may be smaller than thought. His one-sentence view is to keep building quantum computers, but that there is good reason to spend time on the current AI wave.

2025-01-24 · open the original ↗

• AI engineer tops the list of fastest-growing US jobs. The host relays LinkedIn Economic Graph research on millions of jobs started between January 2022 and July 2024, which excluded internships, interim and student roles and jobs dominated by a few employers. He notes it does not disclose its minimum-count threshold. US AI engineers' top skills are LLMs, NLP and PyTorch, and they cluster in San Francisco, New York and Boston. The role is 36% fully remote and 27% hybrid.
• AI consultant and AI researcher also rank among the fastest-growing. AI consultant is second and AI researcher twelfth. The host says the consultant role is less technical (top skill prompt engineering, top prior role operations associate). AI researchers' top skill is deep learning, and only 11% are fully remote and 19% hybrid.
• AI roles are leading growth abroad too. In the reports AI engineer is first in the UK and Netherlands, fifth in Sweden, sixth in Canada and Israel, and twelfth in India. AI researcher is third in Canada and Israel, ahead of AI engineer there, and ninth in the Netherlands.
• The host reads this as data science splitting into specialties. His inference is that data scientist has ceded its high-growth spot to AI-specific titles, not that its skills matter less. He notes data scientist is among the most common roles people move from into AI engineering and research.

2025-01-23 · open the original ↗

• Blogging matters more now, because you own your words. Downie, a principal product manager at Microsoft who maintains dasBlog, says social networks drained blogs through sheer ease, but the risk of losing everything, which he learned when Windows Live Spaces shut down, made owning his words worth the setup. Hanselman calls posting on platforms "digital sharecropping", adding that people now feed AI models unpaid.
• Friction is the barrier, so keep an exit strategy. Hanselman argues that domains and certificates are too much for most people, much as the one extra step of picking a server cost Mastodon users. Downie, who prefers "stick shift" and keeps his blog as plain XML files, says anyone on an easier hosted service needs a way to take their words and images with them.
• The blog should be the origin, not a federated node. Downie considered adding ActivityPub support to dasBlog but gave it up, finding the spec too complex to keep up with even after 20 years of blogging, and prefers to publish on his own site and push elsewhere. Both say putting short posts on a curated long-form blog, as IndieWeb practitioners do, would need a separate notion of short posts.
• Write for your forgetful future self, with full context. Downie's first test is whether a post will help him later, assuming his "goldfish" memory will have lost it; Hanselman's criticism of other bloggers is that they always leave out a paragraph of context. Downie blogs answers to questions he gets repeatedly, including from mentees, as a force multiplier.
• Don't chase audience or cadence. Downie's most popular post, a throwaway on how he writes his yearly review, draws tens of thousands of views, and he blogs about monthly or when the feeling hits. Hanselman says 100 readers is a real audience, urges people to publish their drafts, and says a long-running blog shows more about a technologist than a follower count.

2025-01-23 · open the original ↗

• Tailscale's value is a control plane layered over WireGuard. WireGuard (ChaCha20 encryption, public/private key authentication) carries the data; Tailscale's coordination server tracks every endpoint and builds direct point-to-point links, giving a mesh instead of a hub-and-spoke VPN where remote users talking to each other must cross the main office, adding latency and a single point of failure.
• Devices behind double NAT connect without any port forwarding. Where a direct link fails, as behind Starlink, traffic goes through a DERP relay server using STUN and ICE; the host says the relay only forwards packets and cannot decrypt them, and running your own DERP server is possible but "totally unnecessary".
• The free tier covers a typical household. Tailscale counts machines and users separately, and the free account includes 100 machines; the host's example is three family members across phones, tablets and laptops, and he says he has not outgrown it. MagicDNS names every device as machine name dot tailnet name, avoiding a separately maintained DNS server.
• A pfSense router can share its LAN or internet connection. On a Netgate 4200 behind double NAT, he enrolls the router with a pre-authentication key, advertises 192.168.1.0/24 as a subnet route, and reaches a CasaOS box at 192.168.1.102 from an iPhone on LTE. Advertised subnets and exit nodes both need admin approval in the dashboard; users then pick an exit node to tunnel all their traffic.
• Access controls are powerful but written in JSON. Tailscale denies by default, but its starting rule allows any source to any destination, so most home users need not touch it. Rules are directional yet stateful (return traffic is allowed), can reference users, groups and devices rather than IPs, and are pushed to and enforced on each client.

2025-01-23 · open the original ↗

• Dev containers standardize a whole dev environment in one config file. Ayers says containers, unlike VMs, share a Linux kernel and so are small, fast and immutable; the dev container spec at containers.dev turns that into a config that names a language base (Python, .NET, Java, Go), extensions, ports and tools such as the Azure CLI, supported by VS Code, JetBrains and GitHub. He argues it makes onboarding "a couple of clicks" instead of weeks of requisitioning and installing software.
• Checking the dev container into the repo ends version conflicts. Ayers says the config lives in the repo's .devcontainer folder, so everyone gets the same tool versions, extensions and formatting settings without NVM or virtual-environment workarounds, and VS Code offers to open it automatically. Nearly all his own repos have one, and he cites CPython and Home Assistant as major projects that use them.
• Codespaces is the same dev container, hosted in the cloud. Ayers says opening a repo in a Codespace spins up a dev container and connects browser or local VS Code to it, with port forwarding and debugging, which removes the local dev container's shortcoming of needing Docker and a powerful machine. GitHub accounts get free monthly minutes, and sizes range from 2 to 16 cores, so a Chromebook can drive a 16-core, 64 GB machine for ML work.
• Without a config you get a universal image; with one, full control. Ayers says Codespaces defaults to a universal image with recent Java, Node, Python, .NET, Go and Rust, but will use a repo's own devcontainer.json. VS Code's "add dev container files" command picks a base (Ubuntu, Alpine), layers tools such as kubectl or Terraform, and sets lifecycle hooks on start or rebuild; Dependabot can flag outdated dev container components.
• Start with Microsoft's "try" sample repos and the free docs. Ayers recommends installing the VS Code Dev Containers extension and cloning the "try dev containers" repos (Rust, Java and others), where breakpoints in the container hit from the desktop browser. He points to the VS Code and GitHub Codespaces documentation, containers.dev for the spec and integrations, and JetBrains' own docs.

2025-01-23 · open the original ↗

• Dictating a day's jumbled tasks to a chatbot yields an organised plan. After the school run Azhar records up to 10 minutes of mixed tasks in Claude's audio mode (a speech, car insurance, a parking ticket, a Stripe refund), and on reaching his desk 15 minutes later finds a sorted to-do list, a drafted appeal letter and ordered speech notes. He calls Claude 3.5 Sonnet good enough to change his day.
• The missing piece is turning plans into actions. For now he copies and pastes; he says AI companies are promising that step "this year". He expects hundreds or thousands of agents working for each person and wants a "chief of staff" supervisor agent that takes unstructured input, doubting Apple will build it.
• His multi-agent debate succeeds about half the time. He defines a historian of technology, an investor and a cynic, run on o1 and at least two Claude 3.5 models, with an orchestrator making them argue. A run takes five or six minutes and costs about 20 cents, and fails about half the time, which he says matches human teams on open-ended research. He builds it in Wordware, disclosing he is an investor, or Lindy.
• Learning AI means working close to the API. Azhar says calling models directly lets you set temperature (sensible but creative, around 1, for an evaluator), which third-party interfaces hide. He also uses Gemini Advanced research, NotebookLM, and an email-drafting tool he has invested in; the host describes using You.com's research mode alongside Claude 3.5 Sonnet.

2025-01-22 · open the original ↗

• Who is the user: the user is the hero, you are Yoda. Smith, a software engineer of about three years at Microsoft building an internal design system's component library, says his users are the developers consuming it, not end customers, and that DX and CX are subsets of UX that non-designers wrongly treat as someone else's problem. Teams cast themselves as Luke Skywalker and showcase their product, when their job is to help the user succeed; Giard offers his own early failure of building data-entry forms that mirrored relational tables, so users had to learn referential integrity to delete a customer.
• State what they want and how you help, repeatedly. Smith's second and third questions overlap: users skimming search or ChatGPT results are in "fast survival mode", so any ambiguity in what you solve gets missed and the message has to be broadcast "over and over". Whether they adopt the product, he says, depends on how well you articulate their success story with them at the centre.
• Question four asks what you need back from the user. Smith says it is not the usage instructions but the business outcome the builder needs (a purchase, an email sign-up, clean data, acknowledgement that secures funding), and conflating it with what the user wants leads to frustrated experiences. His example is a grocery store such as Kroger trading loyalty-card discounts for customer data; Giard frames the split as question two being the user's needs and question four the provider's.
• Question five: incentives are sometimes the experience, sometimes explicit. A clean way to solve the problem can be incentive enough, as with Stripe's developer experience, Smith says, but asking for money or personal information usually needs an explicit exchange like the loyalty card.
• Expect to get the answers wrong first, then iterate. Smith says elaborate personas from UX workshops are usually wrong, which is fine if you ship an MVP, learn, and re-ask the questions with each release; low engagement signals a wrong problem or wrong users. Users also prescribe solutions (as in the "very apocryphal" Henry Ford "faster horse" line), and he tells a story of an office that answered slow-elevator complaints by installing mirrors rather than a costly faster motor.

2025-01-22 · open the original ↗

• Public clouds grew out of platforms built for companies' own services. Russinovich traces Azure to the mobile shift, Ray Ozzie's 2005 services memo, and Project Red Dog, formed as he joined Microsoft in 2006, which Dave Cutler joined; AWS launched EC2 and S3 a year after the memo for Amazon's own needs. He says opening a platform to strangers adds most of the work, because you cannot trust the public as you trust your own staff.
• Cloud economics come from pooling and self-service. Enterprises provision for peak load and leave 30% to 60% headroom idle, while a multi-tenant cloud oversubscribes statistically because customers peak at different times, running close to 100% utilization. The other gain is agility, such as Scott Guthrie's early goal of a website in 90 seconds.
• Azure assumes "hostile multi-tenancy", isolated by virtualization. Russinovich coined the term against friendly enterprise multi-tenancy, treating the hypervisor's small trusted computing base as the security boundary, with networks virtualized too. Placement is handled by an allocator he co-authored a paper on (rendered "protein"), which packs VMs to avoid fragmenting servers and uses ML to predict arrivals; he puts the fleet's out-for-repair share at about 3% or lower.
• Azure launched as Windows-only PaaS; sales forced IaaS. Announced at PDC 2008 and commercially launched in February 2010, it required apps rewritten as cloud services; in 2011 Microsoft's head of sales told Amitabh Srivastava he would stop selling it without IaaS. Russinovich and Corey Sanders wrote the plan, IaaS shipped about a year later with Linux VMs, and he says Azure now likely runs more Linux than Windows VMs.
• Most developers need not learn the layers beneath them. Russinovich cites falling sales of the Inside Windows books, from about 250,000 copies for the 1996 edition to 75,000, 50,000 and 40,000 for later ones, as abstraction making internals specialist knowledge. Hanselman's advice is to learn one layer below your comfort zone.

2025-01-22 · open the original ↗

• Linear forecasts ignore how things actually grow. Azhar says business plans still show fixed increments, though growth follows a logistic S-curve with an exponential phase; "nothing is linear except our birthdays". He points to compound interest and late pension saving, and says Bay Area tech internalised Moore's Law. Listeners have lived through exponentials in phone upgrades and data lakes growing from a gigabyte to a petabyte.
• Model growth as compounding percentages with feedback loops. He says he has never used linear increments in more than 20 years of planning, and builds in feedback loops such as Intel chips enabling Windows features that drove upgrades. A linear multi-year S3 storage budget, he argues, is the extreme assumption, even if a realistic one alarms the CFO.

2025-01-21 · open the original ↗

• The host is selling surplus office gear through the Crosstalk store. Prices include shipping: an Anker 8-in-1 MagSafe charging station for $40 (from a $49 grab bag; valued at $69), and a Wyze Cam v3 bundle for $85 (three cameras, two used, plus a spotlight, socket lamp and solar panel; about $163 new).
• An unopened Arista Wi-Fi 6 access point goes cheaper than eBay. The Arista C-200 is a 2x2 MIMO Wi-Fi 6 AP with a single gigabit port, still in its packaging; he puts eBay at about $45 plus shipping and asks $35 shipped. The items are listed under "Tech Sale Tuesdays" in the store on the Crosstalk Solutions website.

2025-01-21 · open the original ↗

• Assuming exponential compute growth stops is the more radical bet. Azhar calculates world computing capacity has grown about 65% a year compounded for 52 or 53 years since 1972, and his baseline for 10 to 30 years out is to extend that line. He grants the turkey-before-Thanksgiving objection but argues the curve rests on processes (parallelism, interconnects, financial incentive, soon AI help) rather than on shrinking transistors alone.
• Linear forecasts are the extreme assumption in planning. He says he has modelled growth as compounding percentages with feedback loops (Intel chips enabling Windows features driving upgrades) for more than 20 years, and that a linear S3 storage budget contradicts history. Most real growth follows a logistic S-curve; "nothing is linear except our birthdays".
• AI-native companies are opening an "exponential gap". Azhar says fast-growing AI companies reach $30 million annualized revenue in about 20 months against about 60 for SaaS, and that Anthropic reached a billion dollars in sales in
  2024. His advice is to build "the muscle of working with AI" on APIs and semi-finished tools, neither finished SaaS workflows nor your own foundation models.
• Energy is turning from a commodity into a cheapening technology. He says fossil energy has not got cheaper in 200 years while solar panel costs fell by, he thinks, 99% over 20 years, with batteries on a similar curve, and likens the shift to telecoms giving way to the internet. He cites an Oxford paper from about two years earlier finding a fully renewable system cheaper to run, and puts the energy transition at a couple of decades and cheap carbon capture and desalination further out.
• AI is needed to keep knowledge growing as population peaks. Azhar expects human population to peak in 60 or 70 years, after which fewer people can work on problems. On jobs he says there is "no solid ground": citing David Autor's finding that 60% of today's US jobs did not exist in 1940, he expects most jobs to change, everyone's work to become more "boss-like" directing AI, and skilling up as the only preparation.

2025-01-20 · open the original ↗

• AWS's generative AI stack runs from plug-and-play to fully granular. The guest (Eremenko, in episode 853) says Amazon Q offers no choice or customisation of the base model. Q Business is effectively RAG across S3, RDS, Kendra, Gmail, Dropbox, Slack, Zendesk, Jira and Salesforce, with an option to fall back on the model's own knowledge. Q Developer is a coding assistant in JetBrains, VS Code and the CLI, and an AWS-account assistant.
• The middle layer suits most; the granular one needs technical staff. He describes Bedrock as pay-as-you-go access to proprietary and open-source models, with prompting, RAG and agents. SageMaker covers the whole machine learning pipeline with finer customisation and deployment, but needs more technical staff or a partner. The guests disclose no Azure or GCP experience, after over two and a half years on AWS.
• A low-code tool outperformed his hand-tuned models, he says. The second guest (de Ponteves) reports that SageMaker Canvas built, trained and tuned a model in a few clicks and five minutes that outperformed every model he had spent hours tuning on his usual benchmark dataset. He also praises Data Wrangler for preprocessing and JumpStart for vision and NLP foundation models beyond LLMs.

2025-01-19 · open the original ↗

• Instruction-based fine-tuning adds an explicit instruction to each example. The guest (de Ponteves, in episode 853) fine-tuned an Amazon Titan model on niche topics with an instruction to answer in two or three words ("a dog" to which pet is more attention-seeking), and says the result gave concise, straight-to-the-point answers. The difference from domain-specific fine-tuning, he says, is an extra dataset column carrying the instruction. The host notes his former company Nebula instead did domain-specific fine-tuning of Llama models with LoRA.

2025-01-18 · open the original ↗

• Start with cost, modality and the customisation the model supports. The guest (Eremenko, in episode 853) says pricing differs between models, that input and output types (text, image, video) rule models in or out at once, and that you need to know which customisation methods you need before checking which models support them.
• Match inference options and latency to how the model is used. He contrasts real-time uses with no delay (an in-game feature, a natural voice conversation) with asynchronous answers in five minutes and background batch jobs.
• Benchmarks, language and size trade performance against price. He notes that leaderboards score the base model before you customise it, that average performance at a lower price can suffice where accuracy is not business-critical, and that small language models are gaining ground.
• Scalability, compliance, licensing and energy are easily overlooked. He asks how a deployment serving 10,000 users grows to 100,000, flags growing AI regulation and model licences, and suggests weighing a provider's chips and energy use in training and inference.
• In practice, dataset format decided the model. The second guest (de Ponteves) says he chose an Amazon Titan model for one fine-tuning experiment and a Llama model for a medical chatbot because each matched his data's format, and mentions Amazon's just-released Nova models.

2025-01-17 · open the original ↗

• The Switch Flex 2.5G PoE is a compact, powered-uplink 2.5G switch. It is $199 MSRP and available from January 16, with eight 2.5 Gbps PoE+ ports and a 10 Gbps uplink on either RJ45 or SFP+; it is a little larger than the host expected, and comes with wall-mount screws, a bracket and a levelling template.
• The PoE it delivers depends on the PoE powering it. Powered through the 10 Gbps RJ45 port, it shares 12 W across its ports when fed PoE+, 37 W on PoE++ and 57 W on what Ubiquiti calls PoE+++. The host lays out the standards: 802.3af up to 12.5 W, 802.3at up to 25.5 W, 802.3bt Type 3 up to 51 W and Type 4 up to 90 W.
• The host thinks "PoE+++" branding will not scale. It is his term for 802.3bt Type 4, used to tell it apart from Type 3; he argues you cannot keep adding pluses, and that "PoE++ Type 3 and Type 4" would have been safer if a later standard ends up called PoE+++.
• Full power and full bandwidth together mean buying more hardware. As of recording, only the Enterprise Campus 24 PoE and Enterprise Campus 48 PoE can feed it PoE+++, and both are marked "coming soon". The alternatives are a $29 PoE++ injector, which limits the uplink to 1 Gbps, or a $79 210 W AC adapter.
• On his own bench, the host accepts a bandwidth bottleneck for now. Replacing his Lite 8 PoE, he feeds it at 2.5 Gbps from a 2.5G PoE++ port on his Pro Max 16 PoE, since his Enterprise XG has no PoE, then chains a $49 Flex Mini 2.5G (five 2.5G ports, 802.3af-powered) off it. It negotiated at 2.5 Gbps, and adopting and setting port VLANs took about ten minutes.

2025-01-17 · open the original ↗

• Intelligence began with physics and chemistry storing information. The host relays Ray Kurzweil's first epoch from The Singularity Is Nearer: atoms formed a few hundred thousand years after the Big Bang, then molecules, with carbon's stable chains allowing elaborate information. Self-replicating RNA, though not alive, let natural selection act. The host adds his own awe that physical constants are fine-tuned enough for this.
• Biological life is the second epoch. Per Kurzweil, RNA learned to make proteins and became trapped in fatty-acid bubbles to form protocells. Selection for more stable membranes and storage, eventually double-helix DNA, produced single-celled and then multicellular organisms several billion years ago.
• Technology, beginning a few thousand years ago, is the fourth epoch. Humans began storing information outside their brains, from clay tablets to solid-state drives, and now machine learning that perceives and reasons. The host forecasts that simply scaling parameters, inference compute and data will keep machines passing humans on harder tasks, and that within a few years scientific discovery may be led more by machines than humans.
• We are entering the fifth epoch, merging brains with computers. Today brain-computer interfaces are implanted mainly for paralysis or motor loss after spinal injury or stroke. Kurzweil expects ordinary brains to be augmented over coming decades, vastly expanding memory and abstract thought.
• The sixth epoch turns the cosmos into a substrate for computation. The host calls it Kurzweil's most speculative idea: intelligence spreads beyond Earth and restructures matter, limited only by the speed of light and quantum mechanics, until individual intelligences merge. He says it may be wild conjecture, but that Kurzweil has been prescient about AI's pace.

2025-01-16 · open the original ↗

• Quantum computing will be hybrid with classical compute and AI. Svore, a technical fellow and VP of advanced quantum development for Azure Quantum, describes sending a subroutine to a quantum processor and feeding its results back, including as training data no classical machine can produce. She says Microsoft has demonstrated an end-to-end workflow using HPC pre-processing, logical qubits and an AI model.
• The best targets are low-input, low-output, compute-heavy problems. She points to chemistry and materials science, where accurate modelling of strongly correlated electrons matters, with aims such as carbon capture and cheaper, cleaner fertilizer catalysts. Hanselman compares the processor to GPUs, FPGAs and NPUs as a complement rather than a replacement.
• Physical qubits are too noisy, so logical qubits are built over them. Today's qubits fail about once per 1,000 operations, against the one in 10 million or 100 million needed; Microsoft's "qubit virtualization" uses error correction (Hanselman likens it to ECC memory). With Atom Computing it made 20 logical qubits from 80 physical ones, then 28 logical qubits that beat 28 physical ones, shown at Ignite.
• "Quantum advantage" means solving useful problems, not academic benchmarks. Svore prefers the term to "quantum supremacy": just over 50 qubits can beat classical machines on an academic problem, but useful science such as quantum magnet models needs at least 100 highly reliable qubits. The roadmap runs to "level two" resilient computing, then about 1,000 logical qubits needing upward of a million physical ones.
• Scaling is limited by control, which topological qubits should ease. Qubits cannot simply be stamped out by the million because controlling them adds noise, including heat from the control lines. Alongside the neutral-atom machine co-built with Atom Computing and work with Quantinuum, Microsoft is building its own topological-qubit machine, which she says is fast, digitally controlled, and small enough for a million to fit on a credit-card chip.

2025-01-16 · open the original ↗

• The foam insert fits a Starlink Mini into a Milwaukee Packout well. It fits "like a glove", with a notch for lifting the dish out and space for cables and accessories, so the original box can go; but the host had to remove his Striker mount first and would like a version sized for it.
• The host likes the idea but calls the price crazy. The insert costs $90 against about $80 for the Packout itself, and he invites anyone who can make such inserts more cheaply to contact him, wanting one under $40 and seeing room for versions for other mounts, DeWalt cases and Pelican cases.

2025-01-16 · open the original ↗

• The first three lifecycle steps are the labs' job, not yours. The guest (Eremenko, in episode 853) says data preparation, architecture selection and pre-training, which costs "hundreds of millions of dollars", are done by companies like Meta, Anthropic, Google and OpenAI. He adds that businesses repeating them would make global warming much worse.
• Fine-tuning narrows a general model to proprietary data. He gives examples of medical journals, legal data, or ten years of customer-service call transcripts. The model learns specific terms and style the way a fluent speaker learns a specialist vocabulary.
• Evaluate against business metrics, not just technical scores. He says technical scores like BLEU, ROUGE and BERTScore are not enough, and a model must also be tested on the business tasks it is meant to do.
• Deployment, monitoring and iteration form an ongoing loop. He explains an endpoint as just a URL with parameters, says models degrade through data and model drift like a car needing maintenance, and says changes in the business, in customer expectations or in legislation send you back through steps four to eight.
• A single AWS service covers most of the lifecycle. The second guest (de Ponteves) says Amazon Bedrock handles most steps except pre-training, with data prep done in other AWS services. His students fine-tuned a Llama model on a Hugging Face medical-terms dataset in "dozens of minutes" and got a chatbot that could discuss advanced terms. The host separately notes that 2017 was the transformer paper, and ChatGPT in 2022 was GPT-3 with RLHF.

2025-01-15 · open the original ↗

• A foundation model is a pre-baked base layer businesses rent and customise. The guest (Eremenko, in episode 853) likens it to a cake's sponge layer, baked by well-funded labs such as Anthropic, Meta and Mistral, to which a business adds its own toppings for its use case. He says the analogy itself came from ChatGPT.
• Foundation models are broader than large language models. The host offers a Venn diagram in which LLMs sit inside foundation models alongside vision models of the kind a Waymo car relies on, and the guest agrees, adding image and video models.

2025-01-14 · open the original ↗

• Businesses should customise foundation models, not build them. Eremenko frames a foundation model as a cake's base layer that big labs bake and businesses rent and top. Of his eight lifecycle steps, the first three (data prep, architecture, pre-training costing "hundreds of millions of dollars") are done for you; a business handles fine-tuning, evaluation, deployment to an endpoint, monitoring for drift, and iteration.
• Model choice turns on many factors, and in practice data format. Eremenko lists twelve: cost, modality, customisation options, inference options, latency, architecture, benchmarks, language, size, scalability, compliance/licensing and environmental impact. De Ponteves says his own picks (an Amazon Titan model, a Llama model) came down to which fit his fine-tuning dataset's format. The host relays Andrew Ng's advice to ignore cost while prototyping; Eremenko objects for enterprises certain to roll out.
• Domain-specific fine-tuning is the business default. Eremenko ranks it the most common training-time method, ahead of instruction-based fine-tuning (which de Ponteves describes as an extra instruction column in the dataset), RLHF, continued pre-training and training from scratch. De Ponteves reports fine-tuning a Llama model on a Hugging Face medical-terms dataset in "dozens of minutes" to get a medical chatbot.
• Inference-time levers give large gains for little effort. The pair cover temperature (de Ponteves's story generator repeated one story word for word at temperature zero), RAG over a vector store (a French-dessert assistant from one PDF), agents that split tasks into steps so a simpler model can beat a bigger one, and prompt templates, which Eremenko warns can still be hijacked by injected instructions.
• Pick an AWS service by how much control you need. Eremenko calls Amazon Q (Business, Developer) plug-and-play with no model choice, Bedrock the pay-as-you-go middle layer, and SageMaker the most granular. They disclose no Azure or GCP experience; their AWS focus follows their course business's certifications. De Ponteves says SageMaker Canvas beat his hand-tuned models on his benchmark dataset in five minutes.