Souk's Notes - Unofficial Summaries
Polis Pandit$America's Bread Problem and My 6 Month Sourdough Journey
- This post includes paid content that is not available publicly. See the original source.
GeoHussarJune 13, 2025
- Israel's strike is one long-planned phase of defanging Iran. Reporting early and with confusion over targets, he says Israel launched a preemptive strike on Tehran aimed at assassinating leadership. He reads it as planned well in advance for a Trump presidency, and expects this is only the opening act, not necessarily regime change.
- The real target is the Iran–Russia–China axis, with Syria first. He argues Assad's fall removed Iran's conduit to Hezbollah and Russia's only regional naval base, while Israel destroyed Syrian munitions infrastructure and seized territory, opening an easier air route to Iran. He expects the strikes to hamper Iranian military aid to Russia and oil supply to China.
- This will not become World War III. He argues Iran's past retaliation did next to no damage, and the US will send equipment but no troops to fight alongside Israel. He closes noting Iran has supposedly been on the verge of a bomb since 1979.
Technology and FriendsScott Hanselman on New Technology and Old Friends [EPISODE 855]
- A podcast turns "pick your brain" requests into a lasting artifact. Hanselman, who started Hanselminutes in April 2007, says a coffee meeting "dissipates like a mist", while an episode persists and can be transcribed and shared. For roughly the last three years he has asked guests whether they are in a good headspace to record. Over a few hundred shows, about five said no and rescheduled.
- AI semantic search let him query his own past opinions. At Build he and Yina Arenas built an AI chat search over his transcripts, covering more than a thousand episodes. Asked how his views on AI evolved, it answered that he has been "consistently cautiously optimistic but concerned about the ethics". He agrees, while noting his questions may have biased it.
- AIs are people-pleasing interns, not fact generators. Hanselman says models are eager to give the answer a leading question implies and can confirm a user's bias, so sources must be checked. He cites a Mark Russinovich demo: telling a model the host won a podcast award gets an ungrounded model to invent ceremony quotes. A web-grounded model that finds no evidence can then be "gaslit" into accepting it.
- Use AI as a research assistant, not as the author. Hanselman's wife, a nurse, was offended when told to have ChatGPT write a coworker's recommendation, because it cheapens it. Giard says an AI draft of a colleague's feedback reminded him of contributions he had forgotten or never knew, which he then wrote in himself. Hanselman endorses gathering facts this way but asks whether letting AIs write and read letters just cuts people out, calling these ethical rather than technical questions. Giard adds that he accepts AI for outlines, grammar and style but keeps the thinking his own.
- Early-career people should work one layer below their usual level. Hanselman's examples: open a laptop and add RAM, use the command line, or write some C if you only write JavaScript. He suggests running a model locally with LM Studio, the AI Toolkit for VS Code or Foundry Local, in airplane mode, and watching whether it uses the CPU or GPU. His barbecue story: an overheated iPhone charged at one watt until he set it on an ice-covered plate and it rose to 25 watts, as shown by a USB-C cable with a wattage screen.
Dr. Jon Krohn896: AI (Probably) Isn’t Taking Your Job (At Least Anytime Soon) — with Jon Krohn @JonKrohnLearns
- Headline cases of AI job loss don't survive the data. The host notes an Oxford paper by Carl Benedikt Frey and Pedro Llanos-Paredes linking automation to falling demand for translators. Yet official US data show employment in interpretation and translation up 7% year over year. Klarna, which boasted of automating customer service, has reversed course; its CEO now says "there will always be a human if you want."
- Young graduates' joblessness predates ChatGPT. The host says their unemployment relative to the US average began rising in 2009, 15 years before ChatGPT. Their actual unemployment rate, around 6%, is low by historical standards.
- Aggregate labour data show no AI-driven collapse. Per the host, the white-collar share of US employment rose slightly over the past year. US unemployment is 4.2% with reasonably strong wage growth, earnings growth in Britain, the euro area and Japan is robust, and the OECD employment rate hit an all-time high in 2024.
- Low adoption and augmentation explain the gap. The host cites an official measure that under 10% of American companies use AI to produce goods and services. Where firms do adopt it, they mostly use it to help workers rather than fire them. He calls this an opportunity for practitioners, with more companies still needing help than already served.
- Collaborate with AI rather than compete against it. The host's own view is that there is no imminent jobs crisis, though AI will change work. He advises experimenting daily with the latest LLMs; his favourite for general work is Anthropic's Claude Opus 4, and he uses OpenAI's deep research for his hardest problems.
Scott HanselmanStartup Mindsets with Earl Valencia
synopsis
Valencia, co-author with Dan Gonzalez of Startup Mindsets, argues that founding a company is a long, lucky, high-risk grind that only makes sense if you are tied to a problem you would work on even if it failed. His book is meant to make would-be founders ask that question before they quit their jobs.
Why, not how. Valencia grew up in the Philippines, studied aerospace engineering and found the startup world at Stanford business school. He says most startup books cover how, not whether. His team asked hundreds of Series B-and-below founders why they started and whether failure would bring regret. Most cited something personal and said no. A mission or impact statement, he argues, is what guides hiring, investor and technology choices through the grind.
Mindset and myths. He defines the startup mindset as long-term thinking plus tolerance of risk for later scale. He concedes that traditional businesses often earn more at lower risk. Silicon Valley's advantage, in his account, is talent, hyper-connectivity and a risk-tolerant culture that accepts equity over cash, ingredients other places could replicate. Asian press coverage, he says, tends to highlight success where he wanted to highlight the journey.
Founders and funding. He separates entrepreneurs from "wantrepreneurs" by bias for action. Taking investors' money, he says, is a three-to-five-year duty. He walks through the funding stages, from pre-seed and seed ("valley of death") to a Series A now needing about $1 million projected run rate and 10–15% monthly growth. Investors back trust and mission-market-founder fit, and domain expertise paired with technology makes a "betable" company. He ends on his chapter 15: it is a marathon, and the market has corrected away from quick acquisitions.
- Not everyone should found a startup; the book asks why, not how. Valencia, co-author of Startup Mindsets with Dan Gonzalez, says plenty of books cover how to build a startup but few ask whether you should, and that keeping AI out of the title was deliberate. In interviews with hundreds of Series B-and-below founders, most said they started over something personal and would not regret it if the company failed.
- A mission statement is what gets founders through the grind. Once dismissive of vision, mission and values as an engineer, Valencia says running his own company showed him they settle the hard calls: a capable hire who doesn't share your values, an investor focused only on returns, technology that may be harmful. Without one, he argues, it is easy to quit in the rough times.
- The startup mindset is long-term thinking plus deliberate risk tolerance. He contrasts a traditional business's steady, linear return with a startup's months or years of losses before hoped-for exponential growth. He stresses that traditional businesses often make more money than tech startups at lower risk, and that founders need "intellectual humility" because success owes much to luck and timing.
- Entrepreneurs show a bias for action; "wantrepreneurs" just pitch ideas. His test is whether someone who pitched him three months ago has anything to show. He calls venture money a responsibility rather than a lottery ticket: selling even 1% of the company means a commitment of at least three to five years.
- Investors back trust and founder-market fit, not just ideas. He relays Eric Schmidt's line from a Stanford venture class that investing happens when "greed overcomes fear", and says investors de-risk by trusting the person and by whether this team fits this problem. He puts today's Series A bar at about $1 million projected annual run rate growing 10–15% a month, and calls pre-seed and seed the "valley of death".
GeoHussarA Certain major power will successfully be able to invade the Philippines #philippines #naval
- The Philippines could repel a Chinese invasion. He says China harasses Filipino fishermen over claims including the Spratly Islands and Scarborough Shoal, but flees when US, Japanese, British or Australian ships approach. He argues an invasion would go worse than one of Taiwan: amphibious assault is hardest, and the Chinese navy last fought maybe 50 years ago.
- Distance and US bases give ample warning and deterrence. The Philippines are about 600 km from China versus about 100 km for Taiwan; at 25–30 knots he puts the transit at 13–16 hours, time for US surveillance to set countermeasures. He says four US bases added in 2023 bring the total to nine, making an invasion likely a declaration of war on the US.
Crosstalk SolutionsHosted UniFi? Try this Easy Set-inform Script Today!
- A one-line script set-informs every unadopted device on a subnet. Hosted UniFi controllers need each local device told where to call home via the SSH set-inform command, which is slow by hand or through HostiFi's GUI at 100 devices. The host points to a script from a Ubiquiti employee that scans the local subnet and does the SSH for all of them.
Stock Tok$19 Companies...Racked and Stacked
- This post includes paid content that is not available publicly. See the original source.
GCastCreating a GitHub Repository [GCast 195]
- A new repository needs an owner and a name unique to that owner. Giard clicks the green New button at the top right of the GitHub portal. He skips the available templates, picks himself as owner from the accounts he has permissions on, and GitHub reports whether the name is available. A description is optional.
- Public repositories are visible to anyone; private ones need granted permission. He makes his demo repository public because it contains nothing secret, and says a public repository should probably also get a license.
- README, .gitignore and license files can be generated at creation. The README is a Markdown file explaining the repository's purpose, contents and how to install and use any code. The .gitignore keeps out files that should not be checked in — binaries, files specific to your own IDE, and files holding passwords or connection strings — and can come from a language template; he picks Python. Licenses range from "100% belongs to me" to free use; he picks MIT.
- The generated files can be edited directly in the portal. He edits the README in the browser and commits it with a commit message, and the change shows on the repository page. He calls the result a starter that you would normally bring down to your own computer to build on and collaborate.
Dr. Jon KrohnHeat-Seeking vs. Truffle-Hunting: Two Winning VC Strategies (with Shaun Johnson)
- Hot consensus deals rarely produce a fund's best returns. Johnson defines heat-seeking as backing deals the venture market already knows and loves, which makes them pricier. He cites an adage that the trends hot in any fund vintage are unlikely to be its alpha returners.
- Non-consensus "truffle hunting" bets are cheaper with bigger upside. These are companies investors don't know, or know but don't believe. Johnson says a portfolio needs both kinds. After ChatGPT in November 2022 everything turned consensus quickly, forcing each fund to choose whether to diversify or lean in. AIX holds both.
AI Analyst Lab5 Steps to Master Effective Visualization | Ep31
synopsis
In this Data Neighbor podcast episode, Matt Harrison, a Python educator and author of Effective Visualization, argues that communication is a data professional's most underrated skill and that a chart is a story to be told rather than a plot type to be produced. The method is CLEAR (color, limited plots, explain, audience, references), a fixed sequence of questions that, Harrison says, will change most plots put through it. Harrison ends on the view that current AI can produce basic charts but not the hand-tuned, back-and-forth work professional ones need, and attributes this to missing training data.
Why an opinionated book. Harrison says theory books such as Tufte's explain color and design without showing the code, while library books list hundreds of Matplotlib features without showing how to tell a story; the book aims to bridge the two. Harrison picked Matplotlib for static charts headed into presentations, books or reports. Having recreated the same plots in Matplotlib, Seaborn, Plotly, Tableau and Excel for a Coursera course, Harrison concluded that no-code tools are fine for exploring but make small finishing tweaks hard or impossible. Curation matters, in Harrison's view, because beginners asking for directions get a path without the warnings.
The CLEAR method. Color means few, deliberate colors with the context muted in gray, and not Matplotlib's cycling defaults. Limited means sticking to line, bar, scatter and histogram plots: Harrison's sample of professional visualization teams fell into five buckets, a catch-all of maps among them, and none used radar plots or treemaps. Explain means a takeaway title, a subtitle and annotations, along the lines of "tell them what you're going to tell them, tell them, then tell them again." Audience means a chart for oneself during exploration differs from one for the CEO. References means a source line stating the data and its changes, so the chart is self-contained and can be checked.
Practice and common mistakes. Harrison says the work is iterative. Matplotlib's default of 10 histogram bins rarely fits, moving bin edges can create misleading "comb effects", and an outlier flattens a default histogram. Fully opaque scatter points blur into a blot, so Harrison sweeps size and opacity to find the clusters. Pandas bar plots of time series turn dates into categories and print every full date. To learn, Harrison recommends deliberate practice: recreate known-good charts from their cited data, study designers' style guides, and get feedback from a professional designer. A host adds that one of the most useful dashboards the hosts saw at Nextdoor was a red-and-green week-over-week Google Sheet.
AI's limits. Harrison has not seen an AI turn a CSV into an Economist-quality chart, and finds models fine on basics but hallucinating beyond them, across Matplotlib, pandas, Polars and XGBoost, because professional work and companies' valuable tabular models are not in the training data. A specialized, opinionated AI fed a style through a large context window might change that later; for now, Harrison says, it is no threat to an educator.
- Color should direct the eye, not cycle through defaults. Matt Harrison, author of Effective Visualization and Effective Pandas, opens the CLEAR method with color: Matplotlib's default of a new color per line makes "spaghetti plots" that professionals avoid. Harrison's example is Our World in Data's chart of health spending against outcomes, with every other country in light gray and only the US highlighted; the test is to close your eyes, open them, and note where your attention lands.
- Limit yourself to the few plot types professionals use. Harrison sampled professional data-visualization teams and found five buckets (line, bar, scatter, histogram, and a catch-all mostly of maps), with no radar plots or treemaps; the book covers only the first four. Harrison calls the constraint opinionated but argues it forces better plots, and that a simple, "irrefutable" chart beats one meant to wow.
- Explain the chart with title, subtitle and annotations. Harrison says plotting libraries treat titles as optional and subtitles as nonexistent, though every professional chart has both: the title states the takeaway, the subtitle expands it, and annotations point at anything odd. A host adds the reason: once a chart circulates without its author, readers draw their own conclusions.
- Design for one specific audience, not everyone. Harrison says aiming to please everyone produces something bland that pleases no one; the plot changes depending on whether it goes to yourself during exploration, to colleagues, to the CEO, or to an external audience.
- References make the chart self-contained and verifiable. Harrison notes no plotting library adds a source line, yet professional charts carry one stating the data and any changes made to it, which gives the chart authority and lets readers check it. Those references also enable Harrison's main practice advice: fetch the cited data and recreate known-good charts, which exposes choices such as putting a time-series axis on the right, where the latest value sits.
Dr. Jon KrohnWhat Makes a Good AI Startup Founder? (with Shaun Johnson)
- Early-stage investing backs people who can pivot fast. Johnson says VCs are not genius market timers. Founders learn and pivot, micro and macro, so AIX backs teams that execute "at the speed of light" until they find the point where their offering resonates with the market.
- Teams are scored on AI nativeness and market savviness. Rarely optimal on both, a team strong in AI but weak commercially gets de-risked with advisors. Right after an inflection like ChatGPT, consensus applications such as AI tutoring favour extremely AI-native teams. As the technology commoditises, as the SaaS stack did, non-consensus market insight matters more. A transformer replacement would restart the cycle.
- Builders who read papers beat a dedicated research co-founder. Johnson rejects the host's three-legged stool of CEO, CTO and AI expert. Drawing on his last team at LILT, he prefers production AI engineers savvy enough to read and integrate new papers over "a PhD that can just sit there and read papers".
Scott HanselmanAI and visualizing multidimensional vectors with Pamela Fox
synopsis
Fox, a principal cloud advocate focusing on Python at Microsoft, walks Hanselman from the intuition behind vector embeddings to how they behave in retrieval systems. She ends on what she thinks ordinary users need: not how vectors work, but whether an answer came from the model's weights or from sources it cites and they can check.
Language and code. Fox says Python went from underdog when she joined Microsoft about three years ago to the first language to get an SDK for new Microsoft technologies, because of AI. A linguistics minor who took compilers alongside syntax, drawing parse trees for both, she says LLMs let her bring language into programming, though she wishes she had paid more attention in linear algebra.
What embeddings are. Fox explains embeddings as vectors whose closeness comes from shared contexts in the training corpus. The dimensions are unnamed and reduced from countless possible descriptors, and Hanselman suggests formality is one of them. The relationships depend on the data and how it was fed in. Her examples are a Reddit username that got its own token in early GPT models, and the later practice of a second pass on cleaner data. One OpenAI model was large enough that the PostgreSQL extension pgvector had to raise its 2,000-number limit.
Who is represented. For a multilingual app, the right model depends on users: Fox favours multilingual models in the US but would pick a specialist for a Chinese-only audience. Nomic's figures show how little some languages get, from 234 million English pairs down to 16,000 for Yoruba. Hanselman raises a University of Venda contact asking when Copilot would speak Venda, and who should own and set the bias of models for indigenous languages.
Retrieval in practice. Fox describes chunking, embedding and storing documents with citation metadata. She recommends hybrid vector-plus-keyword search, because vector search always returns something. Temporal and numeric questions, in her view, should be handled by query filters rather than vector similarity.
What users should know. Fox argues that citations are the visible sign of retrieval, and that the application, not the LLM, searches the web. Hanselman contrasts a Google AI answer that insisted it was 2024 with DuckDuckGo's summaries of the visible results. Fox ties this to transparency guidelines in a Microsoft design toolkit.
- Embeddings place words close together when they appear in similar contexts. Fox, a principal cloud advocate for Python at Microsoft, pictures an "ice cream cone" of vectors in which dog and cat sit near each other because they fill the same slots in sentences. The thousands of dimensions carry no names; word2vec, which she trained herself on a Google News data set, is the kind of model that supports "king minus queen" vector arithmetic.
- What an embedding model treats as similar reflects its builders' choices. With OpenAI's early text-embedding-ada-002, Fox found the nearest words to dog were cat and then God, with drug at sixth, which she infers came from character-level training; newer models fixed that. She also gets better multilingual results typing without accents, which she guesses reflects accents normalized in training. Hanselman frames such fixes as product-requirement decisions that build in bias.
- Training data leaves smaller languages thinly represented. Model cards sometimes disclose the mix, and open weights do not mean known training data. Fox cites the Nomic multilingual embedding model: 234 million training pairs for English against 32,000 for Igbo and 16,000 for Yoruba, out of 1.6 billion in total, drawn from multilingual news.
- RAG should use hybrid search and filters, not vectors alone. Documents are split into chunks of a few paragraphs, each stored with its embedding, page number and file name for citations. Fox always recommends combining vector and full-text search, because vector search always returns a closest result however far off. For numbers or "latest", she adds a pre-processing step that turns the query into a filter, such as the last year, before searching.
- Citations mean RAG; LLMs themselves don't search the web. Fox says the key thing lay users should know is the difference between an answer from a model's weights, the most likely to be inaccurate, and one grounded in retrieved sources they can check. The LLM only suggests a query and synthesizes results, while application code does the searching. She cites transparency as a core design principle for AI systems.
GeoHussarWill the insurrection act of 1807 finally be in invoked? 😱
- The administration wants an excuse to invoke the Insurrection Act. He reports 2,000 National Guardsmen deployed in Los Angeles over anti-ICE protests and says the administration signalled for almost two years it would use the 1807 Act, whose text covers "obstruction to the laws", not just insurrection. It has not been invoked yet.
- Escalation signals are building against protesters. He says Defense Secretary Pete Hegseth has offered to deploy Marines stationed about 100 miles from LA, the vice president calls protesters insurrectionists, and 41 anti-protest bills have passed in 22 states since the start of 2025.
Scott HanselmanAI Code Reviews with CodeRabbit's Howon Lee
synopsis
Lee, a senior software engineer at CodeRabbit, argues that an AI code reviewer earns its place through neutrality, context-building and verification rather than through the LLM itself, which, he insists, nobody yet understands.
Review as politics. Lee says much of CodeRabbit's business comes from teams whose reviews are entangled in feuds. A third-party bot that knows no one cools them, much like anonymous peer review.
More than a wrapper. Lee claims half of each LLM call is context the company assembles, beating a whole-codebase paste. Because code is checkable, a verification agent can catch hallucinations, and dozens of linters settle nitpicks.
What LLMs are. Lee says LLMs remain a "god of the gaps" mystery. He draws on the Chinese room, computational complexity and alchemy, and argues new words are needed for intelligence and truth.
Operating it. CodeRabbit picks models rather than offering a picker, ships new Anthropic models within 90 minutes of release and spends 5–10% of its effort curbing verbosity. Lee adds that about 40% of chats are users thanking the bot, which he defends.
- An AI reviewer defuses office politics by not knowing anyone. Lee, a senior software engineer at CodeRabbit, says a sizeable share of the company's business (he puts it at 10 to 30%) comes from teams where personal feuds play out in code review. He likens the bot to anonymous academic peer review; customers, he says, report less politics, though CodeRabbit still wants humans reviewing.
- CodeRabbit's value is the context it assembles, not the LLM. Lee says only about half of its LLM calls carry the code under review; the rest carry Jira and Linear tickets, a proprietary graph of the codebase and clustering to pick relevant context. He claims this beats pasting a whole codebase into Gemini, and estimates building that context by hand would take him about four hours per review.
- Code can be checked, so hallucinations can be caught. CodeRabbit runs a verification agent to find and "stomp on" hallucinations, which Lee says works for code but not for political speech or philosophy. It also runs 40 auto-configured linters, with 50 more planned, to settle nitpick "holy wars" the way Go's formatter did.
- CodeRabbit picks models itself rather than offering a picker. Lee says it tests new Anthropic models for weeks before release and has pushed them to production 90 minutes after launch, sharing internal evaluations with OpenAI and Anthropic. Users choose models only in on-premises enterprise installs. Taming model verbosity is 5–10% of the job, though he notes verbosity also seems to help on objective tasks like bug finding.
- People thank the bot, and Lee thinks politeness is worthwhile. Of more than 20,000 daily chat conversations, he estimates from open-source ones that about 40% are people thanking it, costing tens of thousands. He argues politeness is an end in itself; Hanselman says he even apologizes to voice assistants so rudeness doesn't become a habit.
Crosstalk SolutionsEtherlighting in a Desktop Switch! Just What You Need!
- The USW Pro XG 8 PoE is all-10 Gbps in desktop form. It has eight 10 Gbps RJ45 ports, all PoE+ with a 155 W PoE budget, two 10 Gbps SFP+ ports, layer 3 capability, internal fans and wall mounts for switch and brick. MSRP is $499, in stock in the US UniFi store as of recording.
- Etherlighting colours each port by speed or by untagged VLAN. The host adopted it over a DAC from his UCG-Fiber (UniFi OS 4.2.12, Network 9.2.84) and showed default speed colours: yellow for fast Ethernet, white for both 1 and 10 Gbps, blue 2.5, orange 5. He re-coloured gigabit; breathing mode can be turned off.
- The host rates $499 a good value against rivals. He found an FS switch at $589 with four SFP+ but 2.5 Gbps, layer 2 only RJ45, and a TP-Link at $442 with two SFP+ but likewise 2.5 Gbps layer 2; he could find no other brand matching the specs.
- The 210 W external power brick is his main complaint. He accepts that building it in would force a rackmount chassis, and notes its own mounting bracket lets it hide under a desk. He also says UniFi's layer 3 differs from the standard implementation, without detail.
Dr. Jon Krohn895: The Future of Enterprise AI: Investor Shaun Johnson Reveals What Actually Works
synopsis
Shaun Johnson, a founding partner of the early-stage fund AIX Ventures, holds that AI's limit in the enterprise is now human behaviour change rather than capability. The fastest value, in his view, comes from AI that fills jobs companies are already trying to hire for. On the investing side, he argues that early-stage AI is a bet on teams that combine AI nativeness with commercial savvy and can pivot fast. He also argues that the scarcity of frontier talent shapes everything from salaries to acquisitions.
Why AIX is built this way. Johnson says AI is the inflection point where the next $50–100 billion companies will come from. He describes early stage as higher-return and less correlated with public markets than growth investing. Practitioners Richard Socher, Chris Manning and Anthony Goldbloom sit in every investment committee meeting beside full-time VCs, because founders build at the intersection of AI and a sector. The host adds Stanford AI Index figures: models 1/200th the size of Google's PaLM now reach about 60% on MMLU, and GPT-3.5-level capability costs up to 900x less per token. He wonders whether hype has entered a trough of disillusionment, and Johnson says that must be judged per application.
Judging founders. Early stage is "a game of people": teams that execute fast and pivot until the offering resonates with the market. Johnson rejects the host's three-founder model (CEO, CTO and AI researcher). He scores teams on AI nativeness and market savviness and de-risks the weaker side. In a consensus phase like the one after ChatGPT, extreme AI-native teams win, as with AI tutoring. As the technology commoditises, as SaaS stacks did, market insight matters more, until a successor to transformers resets things. He prefers production engineers who read papers over a dedicated research PhD.
Portfolio construction. Heat-seeking deals are known and expensive; truffle hunting finds non-consensus founders, where he says alpha usually lies. Crowded markets remain open, as Lovable's reported disruption of Cursor and Windsurf shows. Fund two, which the host says closed at $22 million, skips climate for lack of early-stage AI-climate deals. It focuses on enterprise applications and tech-bio. He says "your fund is made by your winners and your reputation by helping your non-winners."
Enterprise adoption. Johnson describes WorkHelix, which he discloses as an AIX investment, as a company that surfaces task profiles from enterprise data, maps bundles of tasks to generative-AI tools and A/B-tests the result. Behaviour change is the obstacle, he says, which explains rollups, professional-services revenue at AI startups, and CEOs mandating AI. The host argues tech hurdles no longer block adoption, citing cascades where cheaper LLMs defer to costlier ones, and expects enterprises to look similar in three to five years. Johnson expects a spread, citing a 5,000-person firm planning for 3,000 in two to three years. He warns that productivity gains with no revenue effect are hard to measure, and staff resist tools that threaten their jobs, so he would target unfilled job descriptions first.
Careers and talent. For practitioners, Johnson advises advocating for AI now and moving toward human-facing roles later, since robotics needs two or three more breakthroughs. Only single-digit thousands work at the frontier, he says, which drives compensation up and pushes big tech toward talent acquisitions. For investors these acquisitions are a downside hedge, as when one AIX team was bought by Google in about three days. He suggests a top-school AI PhD or a move into labs like FAIR or DeepMind as routes into that group.
- Early-stage AI investing is a bet on AI-native, commercially savvy teams. Johnson, a founding partner of AIX Ventures, says practitioners Richard Socher and Chris Manning sit in every investment committee alongside full-time VCs. Teams are rated on AI nativeness and market savviness, and the weaker side is de-risked with advisors. He expects a shift from AI-native teams winning obvious "consensus" applications after ChatGPT to market-savvy teams with non-consensus insight. He argues builders who also read papers beat a dedicated research PhD on the team.
- Portfolios need both heat-seeking and truffle-hunting bets. Johnson says hot deals the market already loves cost more, and a vintage's hot trends are rarely its best returners. Cheaper non-consensus bets are where big wins come from. He wouldn't steer founders away from crowded "red oceans", citing Lovable reportedly disrupting Cursor and Windsurf. AIX's second fund skipped climate because it found too few early-stage AI-climate deals and "atoms are more expensive than bits". The fund targets enterprise applications and tech-bio instead.
- Behaviour change, not technology, is enterprise AI's bottleneck. Johnson says WorkHelix, an AIX portfolio company he discloses, maps an enterprise's tasks from raw data to generative-AI solutions and proves gains in A/B tests. He reads investor rollups of service businesses, and professional-services line items at AI startups, as responses to that friction; one CEO told him they had to mandate AI every Friday. His fastest route to value is to fill open job requisitions rather than make current staff more productive, since teams resist tools that lead to a reduction in force.
- Enterprise change will be gradual and uneven. The host forecasts superhuman ability on narrow tasks within three to five years, but enterprises largely unchanged because of human friction. Johnson expects a spread: one 5,000-person company told him it would be 3,000 in two to three years, as AI takes over the 30–40% of engineering doing patches. The host reports, at the time of recording, about 6,000 Microsoft layoffs, mostly software engineers. Johnson's career advice is to advocate for AI now, and longer term to move toward human-to-human roles, since humanoid robots are two or three research breakthroughs away.
- Scarce frontier talent drives acqui-hires, which also protect investors. Johnson puts frontier AI practitioners at "single-digit thousands", so compensation is skyrocketing. He sees talent acquisitions as the opportunity cost founders must outrun, and as downside protection for investors; one AIX team went from operating to acquired by Google in about three days. Routes into that group, he says, are moving from engineering into labs like FAIR or DeepMind, or a PhD at schools such as CMU, Berkeley, Stanford, NYU, MIT or Toronto.
Polis Pandit$More Anthony Bourdain, Less Toxic Masculinity
- This post includes paid content that is not available publicly. See the original source.
Dr. Jon KrohnAvery Smith on the #1 Mistake in Data Portfolios (And How to Fix It)
- Overused practice datasets count against a portfolio. Asked by a listener what employers hold against newcomers, Smith names the Titanic, Iris and MNIST datasets. For analysts he adds Google's bike-share "capstone", each done by "probably literally millions of people". A project a few hundred people have done is fine. MNIST digit classification alone won't land a computer-vision job.
- A bare GitHub profile is not a portfolio. Smith admits the claim is deliberately provocative. He argues GitHub is a code repository, messy by default, and that recruiters give a portfolio 10 seconds before granting 60. A sparse or fork-filled profile wastes those seconds. He now thinks even his GitHub Pages site was overkill and suggests a simple builder such as Squarespace or Carrd.
Dr. Jon KrohnThe 2 Portfolio Projects Every Aspiring Data Analyst Should Build (with Avery Smith)
- Choose projects you care about, or they won't get finished. Smith says side projects die for lack of motivation, so they should tie to a hobby or a target employer. One killer project aimed at a company like Meta could be enough on its own.
- One SQL project and one data-viz project are enough to start. Smith calls waiting for five projects before applying a trap: "progress over perfection". More projects only help if they don't eat too much time.
- Personal-data projects motivate but can be costly to collect. His 2021 bootcamp built lessons around personal portfolio projects, such as a Looker Studio dashboard of phone screen time and Python on the Spotify API. He notes that Apple won't export screen time, while downloading a CSV into Tableau is far faster.
- An analytical mindset in a non-data job makes résumé material. A former student who was a delivery driver A/B-tested the texts she sent to get into gated buildings. Smith says that became a résumé line ("A/B testing marketing messages") that made her LinkedIn read like an analyst's.
Dr. Jon Krohn“Resumes Talk, Projects Walk”: Avery Smith’s #1 Career Tip
- A project makes a résumé keyword tangible and flips the interview. The host argues a working demo tied to a real passion matters far more than claiming PyTorch, and says he doesn't see how such a candidate could miss an entry-level job. Smith adds that a project supplied before or during an interview replaces "mystery questions" with questions about your own work.
- Public project work can bypass the interview entirely. Smith says he applied three or four times for a Utah Jazz internship without success. With about 60,000 LinkedIn followers, he then posted a project scraping and visualising the team's season shot data and tagged its analytics staff. He says he got the internship with no interview.
Technology and FriendsLaurent Bugnion on Coordinating Hands on Labs [EPISODE 854]
- The conference's hands-on labs drew heavy repeat attendance. At Build, Bugnion, in Microsoft developer relations, was technical co-lead of the in-person labs workstream. There were 48 unique 75-minute labs on provided devices, each led by one or two speakers with proctors. Each ran twice and the most popular ran a third time: 99 sessions over three and a half days. More than 4,000 lab seats were filled, and the average lab attendee did three labs. He says 75 minutes is long enough to go deep without crowding out the rest of the conference.
- Labs offer access to product experts that home learning cannot. Labs are tied to breakout sessions and keynote announcements so attendees can code on just-released technology. Bugnion says some labs are offered on demand afterward. Being there adds product-group engineers and expert proctors to answer questions.
- Months of preparation and selective vetting come before the labs. Bugnion puts it at three to four months: serious work starts around February for a May conference. Room and budget set the lab count, and an internal call for content drew about 100 proposals. Technical experts, including James Montemagno and Scott Hanselman, helped pick about half. Labs run on a vendor's virtualization platform, Microsoft buys its own Azure subscriptions and licenses, and MVPs help test. Bugnion says few labs had technical problems and credits the testing and Microsoft's near-weekly release cadence, which means labs build on stable products.
- MCP and agent labs were the most popular. A lab on using MCP with function apps was the only one with MCP in its title and got an extra session added on the fly after turning many people away. Bugnion says agent labs generally were in demand, including Copilot Studio labs taken by non-technical coworkers. He notes the product UI changed on the conference's first day, forcing a scramble.
- The next fix is matching attendees to the right lab level. These are ideas only, nothing implemented. Bugnion says the session catalog conveys too little for people to choose well, and some Ignite lab attendees had no GitHub account. He floats shorter, less technical labs (what GitHub is, how to commit and open a pull request) and scenario-driven content over product tours, which he expects to be the trend. Work is already under way on Ignite in San Francisco in November and the AI Tour.
GeoHussarReplying to @alexander64888 did 40,000 eastern Ukrainians fight against the Ukraine military in 2014
- Donbas separatism stems from centuries of Russification. He accepts the viewer's point that tens of thousands of eastern Ukrainians fought Kyiv in 2014, putting the figure nearer 30,000 than 40,000. He attributes it to Luhansk and Donetsk being among Ukraine's most Russified regions, a process predating the Soviet Union in which the Russian Empire suppressed Ukrainian identity, briefly revived during independence from 1917 to 1921.
Mr. GlobalHawk
- Claims that Musk stole the 2024 election need evidence. He rejects left-wing claims that Elon Musk hacked the election via Starlink or satellites, says neither Trump's nor Musk's own words count as evidence given their lack of credibility, and likens believers to the "stop the steal" crowd.
Dr. Jon Krohn894: In Case You Missed It in May 2025 — with Jon Krohn (@JonKrohnLearns)
synopsis
The host's May 2025 recap replays four interview clips. Taken together they argue that AI's enterprise value now depends on fitting tools to specific problems: agents for narrow skills, faster dataframes for real pipelines, VR for training that is scarce or dangerous, and clean data under conversational analytics.
Agents versus RAG. Dell's John Roese says enterprises differ by proprietary data and unique skills. RAG chatbots, behind nearly all of Dell's production projects, unlock the data; agents digitise narrow skills and act autonomously under human supervision. He says this needs a different architecture, with a knowledge graph, memory and tool use.
Polars in production. The authors of Python Polars: The Definitive Guide recount moving Alliander's pipeline off a single 700 GB-RAM machine. One job shrank from about 500 GB to 40 GB, and writing the book while using Polars surfaced gaps they reported upstream.
VR education. Mary Spio says SEEK lets one instructor teach at scale. She says it serves needs like training a million prospective pilots for electric aircraft, CPR for new mothers and intubation practice that hospitals now do on paid people.
Conversations with data. Adverity's Martin Brunthaler places success in data quality, harmonisation and monitoring. He describes a staged pipeline from question classification to validated SQL, and plans to use different models for different steps.
- Agents digitise skills, where RAG chatbots unlock proprietary data. In a clip from episode 887, Dell's global CTO and chief AI officer John Roese names an enterprise's two sources of differentiation: proprietary data and unique skills. He says almost all of Dell's production generative-AI projects tackle the first, via RAG. Agents tackle the second by taking an objective and acting autonomously, with the human "on the loop" rather than in it. He frames them as narrow, geofenced skills like driverless airport trains, far from AGI.
- Polars proved itself through a small benchmarked rewrite. In episode 885, the co-authors of the O'Reilly book Python Polars: The Definitive Guide describe moving a pipeline off pandas and R at Dutch grid operator Alliander. The pipeline ran on a single AWS instance with over 700 GB of RAM. A benchmarked rewrite of one small piece won over a sceptical team, and one job shrank from about 500 GB to 40 GB. They say production use exposed gaps they fed back as issues, such as weighted random sampling.
- VR training can scale scarce teachers and risky practice. In episode 889, SEEK founder Mary Spio says one instructor can teach at scale in VR, against a global teacher shortage she cites as 4.6 million. She says an electric-aircraft client needs about 100,000 pilots soon, so a million must be trained. She adds that the brain forms memories in VR as if the experience were real. Her CPR training for new mothers and Baptist Health's nursing residency aim to build the confidence nurses lack, since the average nurse is 50.
- Conversational analytics depends on harmonised, monitored data. In episode 891, Adverity's Martin Brunthaler says data must be complete and aligned across sources, with consistent types and UTC timestamps. It also needs quality monitors, anomaly detection, a retained raw layer and a data dictionary with lineage. His pipeline classifies the question, prompts a model with context, generates and validates SQL, runs it and summarises. A benchmark team tests responses against expected answers, and he says models are converging, so differentiation lies in how components are composed.
GeoHussarThe Polish elections are a huge Russian victory. #poland #Ukraine #easterneurope #donaldtusk
- Nawrocki's win hands Russia a wedge into Polish politics. He says centrist Rafał Trzaskowski lost to Karol Nawrocki, a historian and amateur boxer with no political experience backed by America's president. He expects a president–prime minister split: Donald Tusk wants Ukraine in NATO and closer EU ties, while Nawrocki wants the opposite.
- The prime minister holds more power, but the president can obstruct. He predicts that if Nawrocki prevails Russia benefits from an economically isolated Poland and militarily isolated Ukraine. The constraint is that the presidency shapes much foreign policy and can veto parliament's legislation.
Crosstalk SolutionsIt's Annoying when Hotels Don't Give you Ethernet.
- The hotel room's access-point Ethernet ports gave no connection. The host plugged his GL.iNet Slate 7 into both ports on an in-room Aruba wall access point and never got a DHCP address; unsure whether it was misconfigured or blocked, he fell back to Wi-Fi as WAN, though he prefers hardwired.
Dr. Jon KrohnWant to Break into a Data Career? Use Avery Smith's "ETSP" Ladder
- Learn data tools in order of demand and ease. Smith's "Every Turtle Swims Past" ladder (Excel, Tableau, SQL, Python) ranks tools on how in demand and how easy they are. Excel is familiar and Tableau is drag-and-drop. SQL is "probably the most in demand" and easier than Python, which comes last for its steep learning curve.
- Getting hired fast beats learning everything first. Smith's philosophy is to get a foot in the door as soon as possible. The field changes constantly, and on-the-job learning is the most relevant and is paid.
- AI speeds coding but doesn't remove the need to program. Smith says AI gets him 60–70% of the way on code, but without programming knowledge you can't see where it is wrong. He sees it as a tool for professionals, not a replacement. The host uses Claude for most daily work and pays $200 a month for OpenAI's deep research, which he says saves him thousands of dollars of time.
AI Analyst LabBreaking Into ML: Proven Path (Even Without Tech Experience) | Ep30
synopsis
In this Data Neighbor podcast episode, two hosts interview Marina, an ML applied scientist at Twitch who came from political science and small-business management. The episode builds a self-study road map into machine learning and lands on a short formula: don't let fear stop you from starting, do a little every day, and keep building things. Courses get you started, projects drawn from your own life carry you forward, and communication is what separates the mentees who advance fast.
The path. Marina studied political science, worked on campaigns and at nonprofits, then ran a two-person jewelry company for four years. A public policy master's that turned out to be mostly social data science, with coding, statistics, causal inference and some ML, deep learning and NLP, led to a year at a Berlin statistical consulting firm, then Coursera, then Twitch. The Berlin job came from cold-emailing 30–40 local firms; one replied and created an internship, which became a junior role after about a year. A host calls this growing your "luck surface area". The master's covered no SQL, applied ML or deployment, so Marina filled the gaps through self-study around full-time work, including roughly 20 Coursera courses during her time there, and now mostly through technical books plus projects. Her channel's "gratitude-driven" name is her alternative to hustle culture.
Learning and promotion strategy. Marina advises fundamentals first, Python and statistics, and not tools like PySpark or Airflow, then talking to practitioners to learn which tools their jobs actually use. On promotions, she says knowing each level's criteria and regularly asking a manager where you stand brings out non-technical gaps, such as moving a business metric. AI does not change the fundamentals, in her view: coding agents still delete tests and call them passing, so supervising them takes security and scalability knowledge. ChatGPT helps as a roadmap generator, quizzer and explainer of formulas.
Projects over courses. Once past the very beginning, Marina says to learn by building: pick a problem you care about, find your own data, and learn each next step when stuck, taking it end to end to deployment. Kaggle is fine only for a first project. A host summarizes this as the loop that also prepares you for interview edge cases.
Breadth, depth and resources. Marina lists the foundations for ML: Python, SQL, basic statistics, intuition-level calculus and linear algebra, traditional ML, and evaluation, then software engineering and MLOps. After that comes depth in one area with some room for curiosity. Her controversial view is to learn ML before the math, since people quit at the math and rarely compute by hand. She recommends Coursera's Machine Learning and Deep Learning specializations, 3Blue1Brown, DeepLearning.AI's math specialization, DataCamp, Chip Huyen's Designing Machine Learning Systems and AI Engineering, and Catherine Nelson's Software Engineering for Data Scientists.
Mindset and AI. Against imposter syndrome she advises being comfortable as a beginner, remembering that others show only highlight reels, and knowing that big-tech staff are not all geniuses. On AI's future she calls it a coin toss between very bad and very good, and chooses optimism since the outcome is beyond her control.
- Learn fundamentals first, then ask practitioners which tools matter. Marina, an ML applied scientist at Twitch and formerly at Coursera, tells beginners to start with Python and statistics and leave PySpark or Airflow for later. Her early mistake was learning whatever tool she heard about without knowing if any job used it; talking to people in analytics, software engineering, data science and ML would have narrowed that, and builds a network along the way.
- AI coding agents raise, not lower, the need for fundamentals. Marina says agents like Replit and Cursor still make "crazy mistakes", such as deleting tests and reporting them passing, so learners might shift from syntax toward security and scalability to supervise them. She uses ChatGPT to draft learning roadmaps, as a study buddy quizzing her on calculus, and to explain math formulas.
- Take courses only at the start, then build your own projects. Marina's advice is to learn enough to get footing, start a project, and learn each next piece when stuck; Kaggle is fine for a first project, but after that pick a problem from your own life, since you find your own data and can judge if it works as the subject expert. An end-to-end build, taken to deployment with something like SageMaker or Docker, is what she says impresses in interviews.
- Go broad on foundations, deep on one area, ML before math. Marina's breadth list is Python, SQL, basic statistics, intuition-level calculus and linear algebra (enough for gradient descent), traditional ML, and metrics and evaluation, then software engineering and MLOps; after that, something like 90% on recommendation systems and 10% on computer vision is legitimate. She calls it controversial to start with ML theory before math, because people quit at the math; her picks include Coursera's Machine Learning and Deep Learning specializations, 3Blue1Brown, DeepLearning.AI's math course, and books by Chip Huyen and Catherine Nelson.
- Manufacture luck, and steer promotions by asking about criteria. Marina got a first job by cold-emailing 30–40 Berlin firms with "I barely speak German"; one replied and created an internship that became a junior data scientist role after about a year. For promotions she advises learning each level's criteria and regularly asking a manager where you stand, since gaps are often non-technical, like moving a business metric or cross-functional work; she names communication the most underrated skill.
Crosstalk SolutionsWhat you Need to Know about the Demise of a Telephony Giant!
- Mitel has spent a decade failing to move to the cloud. The host recounts its 2017 ShoreTel purchase, going private a year later, a 2021 cloud sale and RingCentral UCaaS partnership, a Chapter 11 filing early last year citing over $1.15 billion of debt, then dropping RingCentral after three years for Zoom.
- MiVoice Office 250 loses all support in June 2026. It is already end of life; MiVoice Business is still supported, but the host argues that could change at any time. His reading is that Mitel is in survival mode with no clear long-term strategy, a risk for customers who buy phone systems for 5–10 years.
- The host's remedy is moving to Crosstalk's own VoIP service. He openly pitches it as an alternative to RingCentral or Zoom: month-to-month, no termination fees, no per-user licensing, on-premises or hosted, project-managed porting. Mitel, RingCentral or Zoom customers who mention the video get 10% off Crosstalk hardware through 2025.
Dr. Jon KrohnHow Avery Smith Escaped a Dangerous Lab Job for a Career in Data
- A dangerous, low-paid lab job drove him into data. As a chemical lab technician, Smith got hydrofluoric acid on an ungloved hand and washed it off in time. He decided he wouldn't risk his life for near-minimum wage, and wanted the job of the company's data scientist instead.
- An internal pivot is the best route for career changers. Smith moved into a junior analyst role at the same employer, at 25%, then 50%, then 100% of his time. He argues a known, trusted employee is less of a risk than an outside hire without data experience. He credits the firm's size, under 15 people, and doubts it would have worked at ExxonMobil, where he later worked.
GeoHussarReplying to @Chairman Meow Putin won’t be launching any nuclear weapons anytime soon.
- Nuclear fallout in Ukraine would blow back onto Russia. Of five promised reasons, the first is geography: even a small weapon used in Ukraine would spread fallout and environmental damage to Belarus and key Russian cities, possibly Moscow. He says Russia's nuclear planning targets Central and Western Europe or the US instead.
- Putin doesn't know how compromised Russian military infrastructure is. He calls this the crucial reason, citing Operation Spiderweb, planned for over a year and a half, with one drone base reportedly next to a regional FSB headquarters. He also cites an attack on an elite unit in Vladivostok and the underwater mining of the Crimea bridge, while Russia responds only by hitting civilians.
Stock Tok$Mid-Week Movers: HIMS vs. AMD vs. CRWD
- This post includes paid content that is not available publicly. See the original source.
GeoHussarThere has only been one kingdom/Commonwealth that is successfully invaded and defeated Russia
- Only the Polish-Lithuanian Commonwealth ever invaded and conquered Russia. He disqualifies the Mongol conquest of around 1237, which created the Golden Horde, because Russia did not yet exist; there were only Kievan Rus' principalities. He dates Muscovite Russia as a kingdom to 1547 under Ivan the Terrible, after the Mongols were expelled. The captions cut off before the Polish case itself.