Souk's Notes - Unofficial Summaries
Dr. Jon Krohn899: Landing $200k+ AI Roles: Real Cases from the SuperDataScience Community — with Kirill Eremenko
synopsis
Kirill Eremenko, founder of SuperDataScience and of the podcast, walks the host through five recent conversations with members of his learning platform (names and some details changed). His conclusion is that "AI engineer" is still a blurry role taking shape, and that for now the people landing it pair LLM skills with machine-learning fundamentals and, increasingly, production deployment. Alongside the AI track, he holds that a data scientist who translates insight into business outcomes remains valuable. The host's recurring position is that the anxiety of a fast-moving field is best handled by anchoring on long-term trends. Planned as a five-minute episode, the conversation ran to almost an hour and a half.
Fundamentals still get tested. "Alex", late 20s, got an AI engineer job after many rejections. The technical interview asked how he would build an LLM that answers questions over the company's data. He brainstormed fine-tuning, RAG and feature engineering, helped by one collaborative project on the platform. Then the interviewers asked about regression and classification. The host says the term once meant a frontier-lab researcher and now mostly means someone calling existing LLMs. Such work needs data evaluation and exploratory analysis, not stochastic gradient descent, and a three-billion-parameter Llama fine-tuned with LoRA can outperform Claude 4 on narrow tasks. He sees two viable paths: build commercial products at the abstraction layer, or peel back layers over decades, which he thinks keeps raising one's value. Eremenko's reading of Alex's interview is that employers are not yet ready for candidates who work only at the abstraction layer, because a couple of months of prompting is a low barrier.
The field outruns the learner. "Ben", a process engineer in his early 30s, said in August 2024 that he would be job-ready in five months. Ten months later he feels scattered because job requirements keep shifting, and he is studying around a full-time job and family. Eremenko notes LangChain and LangGraph faded within a year and prompt engineering within eighteen months. The host answers that Python, data structures and algorithms, SQL and ever-cheaper compute are safe bets. He says prompt engineering always looked destined to fade as labs fine-tune models to anticipate intent. Eremenko suggests a 60/40 split between those trends and new things. He told Ben to aim straight for AI rather than go through data science first, and the host agrees, framing LLM engineering as one of many specialisations automation has created. In a digression the host doubts Dario Amodei's claim that 90% of code would be LLM-written by end of 2025, noting that such CEOs are also pitching.
Getting past the applicant pile. "Clara", a senior developer in LA in her mid-40s, has years of pandas data-processing in the medical sector and four machine-learning courses. She is targeting $200,000+ roles but faces literally thousands of applicants per job and AI pre-screening. The host calls in-person networking easily the best route, with collaborative online projects as the remote substitute. Eremenko's advice was to ask local recruiters what employers require, learn just that, and return in a couple of months. He stresses that requirements vary by city and industry. The host cites his episode 856, based on LinkedIn surveys, which found AI researcher roles concentrated in San Francisco and AI consultant roles in New York.
Staying in data science. "David", mid-40s and in the Midwest, is targeting $200,000–250,000. He sees lasting value in the customer-facing data scientist who uses AI but does not build it. Eremenko guesses under 3% of the world understands LLMs, so translation remains needed, and the host lists data visualisation, performant SQL, storytelling and Bayesian statistics as durable skills. David hides graduation dates on LinkedIn against age bias, which the host attributes partly to assumed salary expectations. David also treats back-to-office mandates as reduced competition, citing a multi-billion-dollar IBM deal with an Ohio hospital. Eremenko warns that agentic AI automating junior tasks could leave no mid-level or senior people in ten years.
Deployment as the second half of the role. "Evan", a software consultant on a European island, left the platform after a year. His exit interview reported that LLM and agent work increasingly demands cloud and production skills. Eremenko notes that the deployment week of collaborative projects draws members, and that Ed Donner's survey put production deployment top. He sets out two areas: the science of AI (embeddings, data experiments, model choice, LoRA) and production systems (CI/CD, AWS Lambda and Step Functions, whether to use MCP, monitoring, cost and security). The host stops short of calling deployment compulsory but says software skills broaden the jobs open to a data scientist enormously.
- Employers still test machine-learning fundamentals for LLM roles. Eremenko reports that a member ("Alex", late 20s) landed an AI engineer job after a technical interview that asked how he would build an LLM over the company's data (he brainstormed fine-tuning and RAG), then turned to regression and classification. The host's view: entry-level roles may only call APIs, but senior ones need simpler models and the linear algebra and calculus beneath them. Eremenko reads it as employers hedging, because API-only skills are quick to acquire and so prove little commitment.
- Anchor learning on long-running trends, not this year's framework. A process engineer ("Ben", early 30s) planned in August 2024 to be job-ready in five months, and now says requirements shift faster than he can keep up. The host names Python, data structures and algorithms, SQL and falling compute cost as safe ground, and calls prompt engineering an obvious casualty of frontier labs tuning models to anticipate intent. Eremenko proposes splitting study roughly 60% on such trends and 40% on new things, and advised Ben to go straight to AI without becoming a data scientist first.
- Thousands of applicants per job make networking the way in. A senior Python developer in LA ("Clara", mid-40s, targeting $200,000+) finds even her experience doesn't get past AI-screened applications. The host calls in-person networking (meetups) easily the best route. Eremenko advised working backwards: ask local recruiters which boxes employers need ticked, learn only those, and let the recruiter's commission do the job search. The host cites LinkedIn data from his episode 856 for regional variation: AI researcher roles concentrate in San Francisco, AI consultant roles in New York.
- Translating data into business outcomes stays valuable outside AI engineering. A data scientist ("David", mid-40s, targeting $200,000–250,000) is staying out of AI engineering; Eremenko reckons perhaps under 3% of people understand LLMs, so the bridge to non-technical audiences persists. David hides graduation dates on LinkedIn against age bias, and sees back-to-office mandates as less competition locally. He predicts AI jobs near new offices, citing a multi-billion-dollar IBM deal with an Ohio hospital.
- An AI engineer needs both the science of AI and production deployment. A departing member ("Evan") found clients and employers increasingly demand cloud and deployment skills, and Eremenko says Ed Donner's student survey put production deployment top. Eremenko's model: one area is embeddings, data experiments and model choice down to LoRA; the other is CI/CD, AWS Lambda and Step Functions, deciding whether to use MCP, and monitoring, cost and security.
Crosstalk SolutionsSTP Explained: How You Can Prevent Network Loops!
- STP exists to stop loops that cause broadcast storms. Dave Barger, presenting for Crosstalk, says a looped switch can bring a network down; his on-site tip is that link lights all blinking in perfect unison likely mean a broadcast storm. STP breaks the loop with a virtual "STP block" and re-negotiates if a link fails, adding simple redundancy.
- Set STP priority yourself so the right switch becomes root. The switch with the lowest priority (set in multiples of 4096) is elected root bridge; with equal defaults the election falls to MAC address and is effectively random. Barger says to give the switch nearest the router the lowest priority, so the block lands far from it.
- Port cost by link speed decides where the block goes. After the root election each switch adds a speed-based cost per hop to reach the root; in his three-switch example with all 1 Gbps links and the root at 4096, the block falls between switches B and A. Small businesses rarely need to touch costs, only priority.
- RSTP converges faster; deliberate loop designs need other protocols. RSTP adds faster negotiation and edge, alternate and backup port roles, and MSTP handles multiple loops. Barger says clients who plan many intentional loops expecting STP to cope are better served by another switching protocol or dynamic routing such as OSPF.
GeoHussarAnother World War III deterrent is that Russia and China got spooked because our American fighters
- The undetected B-2 raid on Iran should deter Russia and China. He says seven American B-2 bombers flew 18 hours with midair refuelling, went undetected, and struck a hardened target. He infers that Russia and China, whose radar networks are far larger than Iran's, now know the US could reach them undetected with a very powerful bomb.
Dr. Jon Krohn5 Must-Know Tips for Leading AI Change at Work (with Diane Hare)
- Change must run top down and bottom up at once. Hare says senior leaders must be aligned and frontline staff enabled. The change leader's job is often to be the bridge between the two.
- Back every bold claim with proof points. She argues that big storytelling claims cut through noise, but not everyone will believe or follow them. Data and evidence let different populations understand and accept the case.
- Inspire, not just incentivise. Hare says companies double down on incentives such as bonuses and forget inspiration. Describing the impact, and even the ROI, as a cause bigger than oneself taps discretionary effort: the nights and weekends people give because they want to.
- Focus on early adopters, not naysayers. Her fourth tip is to concentrate on "the 18%" of early adopters rather than spending resources on resisters.
- Leaders must have the courage to go first. Hare says change leaders are usually outliers pushing against complacency and "the way we've always done business". She warns they should expect a vulnerable position going in.
Technology and FriendsJeremy Miller on Lessons Learned from Building a Public Open-Source Tool [EPISODE 856 ]
- Contact with real users was his failed framework's biggest lesson. Miller calls FubuMVC, his early-2010s alternative .NET web framework, a "colossal failure" that cost "a couple man years". He blames too little contact with its users. His company, Jasper FX Software, now builds the "Critter Stack": Marten, a PostgreSQL-backed event store and document database, and Wolverine, a messaging, mediator and HTTP framework he pitches as a "clean architecture killer". Working with customers' codebases showed him the friction points. Functional-programming-minded users pushed Wolverine toward plain handler functions and pure business logic with no repositories. The rise of modular monoliths forced changes against design choices made 10 years ago. He says the next big work, handling concurrency (parallel across groups, sequential within one, like an invoice), comes from the problems users raise today.
- Write documentation from day one and keep it mechanically in sync. Early StructureMap docs went stale when he replaced its fluent API, and he says people were angry at him "for years". Marten, started about 10 years ago, had docs from the outset. Its code samples are pulled into markdown from real sample or test code by the MarkdownSnippets tool, so they track the API. He now structures docs around the problems users are solving rather than the framework's own terms. He is borrowing the "duplex book" format of the old Martin Fowler books: flowing tutorials that link out to detailed pages.
- Answer angry users with "how do you wish it was?" Miller says users usually arrive on Discord or GitHub already frustrated. Asking how they would phrase or design it sometimes yields a pull request of "two or three new sentences", and those add up. Every docs page has a link straight to its markdown file on GitHub so readers can suggest edits.
- A tool should "just work" on day one with sensible defaults. A former colleague's "time to login screen" metric measures how fast a newcomer gets from a clean clone to a running app. FubuMVC failed it: infinitely configurable, and likened in feedback to Bowser, "fast, but really slow to get going". Marten, given an open PostgreSQL database, checks its tables and columns and applies the schema delta on first use. Conventions such as treating an
Idproperty as the identity can be overridden. - Short, pointed exception messages are a major usability gain. Miller says FubuMVC's long middleware call stack buried the real error under framework noise. Wolverine's short call stack points straight at the middleware or the user's code that threw. He calls this one of its biggest advantages over FubuMVC.
Dr. Jon KrohnWhy Culture, Not Tech, Blocks AI Transformation (with Diane Hare)
- Merging digital portfolios is technical and cultural at once. Hare describes integrating two medical-technology companies' digital portfolios (systems, apps, data flows) to launch a product. The data was regulated patient data, and the firms ran different stacks, Google against Microsoft, with different cultures, geographies and expertise. Deciding what had to be integrated now and what could wait was the core question.
- Culture's intangibles, not the technology, decide integration success. Hare says power dynamics, how ideas are sourced and how people communicate are what matter, and that they differ with every client. She says her firm otherwise relies on clients' technical staff for the technical side, and values having the host, her business partner, as a technical adviser.
GeoHussarNo, still no chance of World War III. The US just did the hard part for Iran’s enemy #warfare #usa
- No rival can match US-NATO war production, so no world war. He defines a world war as one waged by an enemy able to match its opponent's scale of warfare, and argues neither Iran, Russia nor China can match the United States plus NATO.
- US strikes ended Iran's nuclear option and its bargaining chip. He reports 12 GBU-57A/B bunker busters, said to penetrate about 60–70 m, plus 30 Tomahawks hit Fordow and two other sites. He says Iran has lost almost 70% of launcher capacity and 60–70% of its missiles; Saudi Arabia's nuclear regulator reported no radioactive traces.
- Iran cannot realistically close Hormuz against the US Navy. Iran's leadership vows retaliation against US forces and threatens to close the Strait of Hormuz, but he argues a state that could not defend its most fortified sites cannot hold the strait. He says Russia has left Iran on its own by abandoning their January strategic agreement, and China only wants the oil flowing.
Dr. Jon KrohnWhy AI Adoption Starts Before the Tool Is Ready (with Diane Hare)
- Adoption happens before the tool is ready. Hare says the subject-matter experts whose knowledge a tool will capture should be involved early, through selected champions and early adopters rather than the whole team. In her lab example the case to them was speed of innovation: other scientists needn't retry sequences that failed. The experts' own work also goes faster.
- Spend change effort on the 18%, not the naysayers. Hare credits Simon Sinek with the idea of finding the real early adopters, equipping and celebrating them, and letting them create the tipping point. Resisters, she says, come along when they're ready.
Stock Tok$Some Stocks To Review Over The Weekend
- This post includes paid content that is not available publicly. See the original source.
Dr. Jon Krohn898: My Four-Hour Agentic AI Workshop is Live and 100% Free — with Jon Krohn (@JonKrohnLearns)
- The host's full agentic AI workshop is now free. The four-hour session has him teaching the theory and Ed Donner the hands-on parts. It was filmed at ODSC East in Boston a few weeks before the episode, to a standing-room-only hall, and is posted ad-free on the host's channel.
- The workshop builds up to a multi-agent software team. Module one defines agents using the OpenAI Agents SDK, module two covers agent design with CrewAI, and module three connects agents to tools via Anthropic's MCP. The result is a team of agents in software-development roles that writes an app letting agents trade stocks on real market data in simulation. Module four looks at the future of work, including data scientists' work.
Crosstalk SolutionsUniFi Passkeys Make Logging In Super Easy!
- UniFi accounts now support passkeys for multifactor authentication. The host says it is enabled under Security at account.ui.com, took him under a minute, and stores the passkey in his password manager (1Password) across devices; he keeps TOTP as backup and recommends passkeys wherever possible as more secure and easier than other MFA.
Dr. Jon KrohnDriving AI Adoption Through Simon Sinek's Storytelling Framework (with Diane Hare)
- Original content sticks while recycled content gets tuned out. Hare says the brain files repeated material as already known, a growing problem amid generative-AI content. Employees flooded with email retain perhaps 10% of what corporate communications send them, so change programmes need creative, original stories aimed at "smart, busy people".
- Lead with why, not with what. Hare learned Simon Sinek's golden circle (why, then how, then what) in a three-year partnership while at Ernst & Young. She says most enterprises, and AI projects especially, run it backwards, "trying to find a nail after you have a hammer". Leading with purpose filters for the people who share the cause.
GeoHussarTurkey 🇹🇷 is now the most powerful country in the Middle East. #turkey🇹🇷 #military #nato
- NATO membership makes Turkey the Middle East's strongest power. With Iran decimated, he argues Turkey's edge over Iran's opponent is membership of a 32-nation alliance no country can stand against, plus experience operating inside it. He grants Egypt has the region's largest army and air force but calls its technology obsolete next to Turkey's NATO equipment.
- Turkey ranks second in NATO behind only the United States. He cites a standing army of about 445,000, roughly 1,421 advanced military drones, and about 1,000 advanced ballistic missiles, giving Turkey NATO's second-largest army, drone fleet and air force. He says Turkey aims to produce its own medium- and long-range missiles by 2026.
AI Analyst LabWill AI Take Your Job? A Chief Data Officer Explains - Ep32
synopsis
In this Data Neighbor podcast episode the hosts interview Angi's first chief data officer, earlier a data science and engineering leader at Twitter and Microsoft. The conversation moves from what the role demands to where AI agents are heading, and ends on the guest's position: no one can predict AI's effect on jobs, so the only lever is where you spend your time and energy, and the right stance is to respond to uncertainty rather than react to it.
The role. Angi combined Angie's List, HomeAdvisor and Handy between 2017 and 2019, and the guest's first task was to unify the scattered data teams and bring ML matching to its homeowner–pro marketplace. The difference from big tech, the guest says, is ownership: in place of one functional sliver, everything under the data umbrella is yours, and you sit in leadership meetings on topics far from your expertise. The guest's rule there is to speak on what you know, stay quiet on what you don't, and voice intuition otherwise.
Strategy and execution. Data strategy starts from the customer: the merger left nine to ten unconnected data sources behind one customer, so the work is fixing quality, cutting fragmentation, and explaining concentric priorities to leaders who each see only their function. For speed, the guest imposes an agile, OKR-tracked operating system and a shared warehouse. On precision, the guest argues that each added "nine" of data quality costs more, as uptime does, though executives rarely see it that way, and that because ML success rates run around 10% or lower, a better team delivers more bad news. A host adds that Snowflake Summit vendors demoed agentic tools on flawless data and skipped data readiness.
AI first and the agent future. The guest treats "AI first" against "AI forward", and Anthropic's augmentation-versus-automation terms, as points on a 0-to-1 scale. Microsoft's frontier-firm report stages this from chat assistance to agents managing agents. Joining Sam Altman's wish for a model that improves AI research to Jared Kaplan's scaling law on agent working time, the guest infers agents will come to supervise and improve agents, and expects big companies to buy startups with outsized revenue per employee.
Labor and preparation. The guest declines to forecast labor effects beyond a rough guess that automating about 10% of GDP would bring real disruption. Companies need training and ways to keep people motivated, and individuals should get good with tools such as Cursor. Responding to a host's framing of Anthropic's CEO predicting 50% unemployment, the guest says such forecasts are one possibility among several. The closing advice on talking to executives is to answer first and see the problem through their eyes.
- A chief data officer owns the whole data body, not a slice. The guest, Angi's first chief data officer and earlier a data science and engineering leader at Twitter and Microsoft, contrasts the "sliver" a big-tech leader owns with a role where "the buck stops with you"; at Angi, formed from Angie's List, HomeAdvisor and Handy between 2017 and 2019, the first job was to pull dispersed data teams into one organization and bring ML matching to the two-sided homeowner–pro marketplace. In leadership meetings the guest advises speaking plainly on what you know, staying silent on what you don't, and voicing intuition ("spider senses") in between so one-way-door decisions don't rest on a single voice.
- Data strategy works backward from the customer and cuts fragmentation. Three merged companies meant nine to ten data sources that did not talk to each other behind one customer; the guest's strategy asks which sources must improve to enable customer scenarios, minimizes tool and source fragmentation, and orders priorities as concentric circles that business leaders, each seeing only their own function, must be walked through.
- Speed comes from standardizing, and every extra "nine" costs more. The guest runs data teams on an engineering-style "operating system" (agile cadences tied to quarterly OKRs) and standardizes on a central warehouse such as Snowflake or Databricks so teams move at one velocity. People accept that 99.99% uptime costs far more than 99%, the guest says, but not the same for data availability, cleanliness and freshness; and since a strong ML team's success rate is around 10% or lower, getting better means delivering more bad or inconclusive news.
- "AI first" versus "AI forward" is a continuum, not a binary. The guest places it between 0 and 1, from augmentation to automation, drawing on Microsoft's "frontier firm" report, whose stages run from chat-tool help to humans managing agents at some optimal human-to-agent ratio to agents managing agents. From Sam Altman's and Jared Kaplan's remarks at Snowflake Summit (a new scaling law on how long agents work alone, publicly about an hour now), the guest infers self-improving agents are being planned, and expects large companies to acquire startups claiming 6–10x revenue per employee.
- Respond to AI uncertainty with agency rather than resignation. The guest says no one knows the labor impact, offering 10% of GDP automated only as a thrown-out threshold for serious disruption, and that companies lack training and a way to keep motivated the people whose work is being automated. The advice is to join the shift (a senior developer learning Cursor), since tuned prompts become a competitive edge, and with executives to answer first, then walk them from their simplified frame to the reality.
Dr. Jon KrohnThe Biggest Mistake Enterprises Make with AI Rollouts (with Diane Hare)
- AI initiatives fail when siloed from current enterprise priorities. Hare says product and data teams may find the best use cases, but they can't get them adopted without senior leadership buy-in. They also need a roadmap in concert with what the enterprise prioritises that quarter. Otherwise the project feels like a "foreign, non-priority siloed mission".
- Scoring use cases is only the first step. In her sequence the host's firm estimates ROI and the use cases are scored down to a top five. Then come socialisation, alignment and approval with leadership, and an assessment of change impact and readiness. The two are business partners, and the host says the steps apply whether or not either firm is involved.
Crosstalk SolutionsNonprofit Network Design for under $1000 bucks!
- The sanctuary's own plan meant double NAT and a weak mesh. The Yellowstone Wildlife Sanctuary, a nonprofit with a $1,000 cap, had bought two Dream Router 7s and two U7 Outdoors, planning one UDR7 behind its ASUS router and a wireless mesh of about 80 m (250+ ft) to a barn; the host rejects both.
- One UDR7 replaces the ASUS as the dual-WAN gateway. Its second WAN leaves room for Starlink, which the sanctuary wants but cannot yet afford monthly. Three USW Lite 8 PoE switches add ports in the office, gift shop and barn, with a U7 Outdoor on omni antennas at the shop and the barn.
- A NanoBeam point-to-point bridge replaces the mesh to the barn. A pair of NanoBeam 5AC Gen 2 at $99 each should carry about 300–400 Mbps over that span. They need 24 V passive PoE, so an INS-3AF-I-G converts the switch's 802.3af output rather than using an injector.
- Surge protectors at both ends guard against lightning. For the Montana weather, a grounded ETH-SP-G2 goes inside where the cable exits and another outside near each NanoBeam, so a strike should kill only the radio and protectors. Protectors for the U7 Outdoors, under the eaves, are optional.
- The new gear totals $655, under the $1,000 budget. That is three switches at $109, two PoE converters at $20, four surge protectors at $12.50, two NanoBeams at $99 and two mounting arms at about $20 from Amazon. The spare UDR7 could resell for an estimated $200–250. The design was worked out on a live stream with viewers.
GeoHussarIran is losing its missile capabilities and it’s running out of options. #iran #defense #war #news
- Iran's shrinking barrages show its launchers are being destroyed. He counts about 200 missiles on day two, 100 on day three, 50 on day four, and only 10 so far today despite a promised record barrage. He infers satellite tracking of each launch located the launchers, and says Iran has lost 35–40% of them, so a 3,000–4,000 missile stockpile matters little.
Dr. Jon KrohnHow to Enable Enterprise AI Transformation, with Strategy Consultant — with Diane Hare
synopsis
Diane Hare, a strategy and storytelling consultant and the host's new business partner, argues that enterprise AI stalls on people and alignment, not technology. In the arrangement disclosed on air, his consultancy implements and hers handles change enablement. Her answer is to tie AI projects to what leadership already prioritises, and to win adoption through original, inspiring narratives aimed at early adopters. She ends with five rules for change leaders.
The problem. The host opens with CrewAI COO Rob Bailey's line from ODSC East that tech isn't the barrier to multi-agent systems. He doubts the cause is job fear; he thinks it is the difficulty of agreeing what to prioritise when better tooling may arrive in six months. Hare adds that most enterprises lack clean data and skilled practitioners: they want "the Ferrari" without "the highway" of governance and structure.
Prioritising. The host offers RICE scoring (reach times impact times confidence, divided by effort) as more nuanced than ROI alone. Hare notes that only product or data teams have time to score long use-case lists. Their best ideas still fail without senior buy-in and a roadmap in concert with that quarter's enterprise priorities.
Storytelling. Hare says original content sticks while recycled content is tuned out, and that staff retain about 10% of corporate communications. She learned the craft in a three-year Ernst & Young partnership with Simon Sinek. His golden circle runs why, how, what, and she says most enterprises invert it.
Case studies. For upskilling, she builds separate curricula for executives, middle managers and frontline staff, sometimes on interactive platforms built with Ceros. Her own firm permits AI for discovery but not for writing drafts. With lab scientists reluctant to share formulas, she framed the tool as speeding breakthroughs and studied the bench workflow. She also enlisted champions before launch, saying adoption "happens before the tool is ready", and focused on the 18% of early adopters. In a medtech merger of two companies on different stacks and cultures, she says the intangibles, not the technology, decided the integration.
Five tips. Her closing rules are: work top down and bottom up; back bold claims with proof points; inspire rather than only incentivise; focus on the 18%; and accept that leaders go first.
- People and readiness, not technology, block enterprise AI. The host quotes CrewAI's COO Rob Bailey at ODSC East: tech isn't the barrier to multi-agent systems. He places the difficulty in getting an organisation to agree what to prioritise, rather than in fear of job loss. Hare adds that most enterprises lack even clean data and trained practitioners: "we want the Ferrari, but we don't even have the highway."
- Good use cases die without leadership buy-in and enterprise fit. The host ranks ideas by RICE score (reach × impact × confidence ÷ effort). Hare says product or data teams can produce that list, but adoption needs a roadmap that fits what the enterprise prioritises that quarter. Otherwise the project reads as a "foreign, siloed mission". Her sequence is to score, take a top five, then socialise, align and approve, and then assess the enterprise's readiness to change. The two are disclosed business partners: his firm implements and hers handles enablement.
- Original stories cut through corporate noise. Hare says recycled content gets tuned out and employees retain perhaps 10% of corporate communications. She credits a three-year partnership with Simon Sinek at Ernst & Young for her method: his golden circle of why, then how, then what. She argues most enterprises run it backwards, leading with what, "a nail after you have a hammer".
- Adoption is won before the tool ships, with early adopters. In a lab case, scientists resisted putting formulas into a tool that would share their expertise. Hare's answer was to inspire rather than only incentivise, framing the tool around speed to breakthrough, and to observe the lab workflow (gloves, safety rules). She recruited champions early and opened two-way feedback with the tech team. Her closing five tips: top down and bottom up; bold claims backed by proof points; inspire or incentivise; focus on the 18% of early adopters, not naysayers; and lead by going first.
- AI upskilling needs separate curricula by employee type. Hare says executives, middle managers and frontline or manufacturing staff each need their own track. In one offsite, a global HR team had to upskill itself and the wider company. Her own storytelling firm first banned AI for writing; it now allows it for discovery and input, but drafts must come from the writer.
Crosstalk SolutionsInternet Anywhere with this Starlink Mini Travel Setup
- A Starlink Mini kept the host online where LTE failed. On a 7-hour road trip it worked lying on the shelf behind the rear seat, powered by an Anker power bank showing 22 W while searching, about 10–12 W once connected, and an estimated 4 hours 13 minutes of runtime. He pitches his referral link for one free month of service.
Stock Tok$Put My Money Where My Mouth Is 😮
- This post includes paid content that is not available publicly. See the original source.
GeoHussarIran has signed that they want to de-escalate hostilities. But they are still being bombarded
- Iran is seeking a face-saving exit from the war. Citing IranWire, an anonymous Iranian diplomat said Iran would abandon uranium enrichment but wants a face-saving way out and guarantees the US stays out of the fight. He says Iranian barrages have damaged civilian infrastructure but done little to military targets.
- Khamenei's fate is unclear amid a successful decapitation campaign. He says Iran's opponent removed the bulk of Iran's military leadership, with reports the US blocked an assassination of Khamenei. Other reports suggest he may be dead, which he flags as possible propaganda; the most official claim has him in a bunker near Tehran with his family.
- Falling oil prices signal tensions are easing. He calls this the number-one indicator, citing oil at $68.9 a barrel.
Polis Pandit$Liberation Day’s Lasting Impact: How Treasury Markets Check Executive Power
- This post includes paid content that is not available publicly. See the original source.
Dr. Jon KrohnNew Opportunities for AI Startups in 2025 (with Shaun Johnson)
- Crowded "red ocean" markets still reward insightful founders. Johnson says opportunities exist across the spectrum from consensus to non-consensus applications, as they did in SaaS five years ago. He disputes investors who call AI coding too crowded with Cursor, Windsurf and Cognition. Incumbents that grow big become ripe for disruption, he says, citing Lovable reportedly disrupting Cursor and Windsurf, and AIX is about to back a team entering a red ocean.
- Deep, specific domain backgrounds make overlooked verticals compelling. He cites a founder innovating in one narrow area of consumer-packaged-goods product development. That founder's background matched the problem unusually closely.
GeoHussarThis is not World War III, we are far from it. It’s just a repositioning of power Towards the west
- The Iran war is lopsided and regional, not World War III. He says Iranian missiles and drones got through to hit apartment blocks and perhaps one federal building, while Iran's opponent killed over 20 top commanders and six or seven nuclear scientists and struck nuclear, military and energy sites, with the Tehran refinery ablaze.
- Iran will lose if the war continues. He says Iran's air defences have been all but dismantled since last year, making billions spent on Russian S-300 and S-400 systems worthless. After six or seven barrages, Iran has not killed a single member of its enemy's military leadership.
Crosstalk SolutionsGiveaway! Win This Alta Labs NFR Kit!
- NFR kits give installers the margin MSRP pricing leaves out. The host says installers buying at MSRP end up paying for their own time on warranty and RMA work; a not-for-resale kit lets prospective resellers buy and learn a vendor's gear at a steep discount.
- Alta Labs' kit costs $399 against a $636 MSRP. It includes the Route10 gateway (four 2.5 Gbps RJ45, two PoE, two 10 Gbps SFP+), the S8-POE switch (8 gigabit ports, four PoE+), an AP6 (2x2 Wi-Fi 6) and AP6 Pro (4x4). It requires joining the reseller program; code "crosstalk" takes $10 off. The host is giving one away, winner picked June 27.
GeoHussarIran managed to strike into the capital of their enemy. But only 7 to 9 out of the 150 missiles
- Iran's strikes hurt civilians while its own military took the blow. He says only 7–9 of over 150 Iranian missiles hit, so almost 95% were repelled; using a ChatGPT-compiled list he says he verified, the toll was one civilian killed and dozens injured or hospitalised. Iran's opponent hit 150 sites across Iran, destroying air bases and nuclear testing facilities, apparently closing Tehran's airport and killing six nuclear scientists and 20 senior commanders.
Dr. Jon KrohnHow to Future-Proof Your Career in the Age of AI (with Shaun Johnson)
- In the short term, become an advocate for AI. The host frames the question with roughly 6,000 Microsoft layoffs, reportedly mostly software engineers, in the week before recording. Johnson's first advice is to treat AI as a thinking partner that helps generate solutions, not a passive tool like a spreadsheet macro.
- Longer term, move toward human-to-human touch points. Johnson says AIX expects digital work to be transformed while robotics stays far off. In his view humanoid robots need two or three major research breakthroughs. Roles in outbound sales, client work and contact with senior leadership will therefore stay valuable for a long time.
GCastDeleting a GitHub Repository [GCast 196]
- Deleting a repository is deliberately hidden in its Settings. From the repo's Settings button, the option sits at the very bottom in the "Danger Zone", where irreversible actions are outlined in red with red buttons. Giard says he made the video because the feature is not at all obvious, and judges it hidden by design so nobody deletes a repository by accident.
- GitHub stacks several confirmations before the delete goes through. After "Delete this repository", it asks him to confirm the repository, then warns that deletion is permanent with no undelete option, then asks him to type the owner/repository name. He shows that the name can be copied and pasted rather than typed.
- With two-factor authentication on, a final approval comes from the phone. Because his account uses two-factor authentication, GitHub asks him to approve in the GitHub Mobile app using a number shown on screen, which changes every time; entering a password is offered as an alternative. Once he approves, the portal shows a message that the repository was deleted.
Dr. Jon KrohnWhy Enterprises Struggle to Implement AI — and How to Fix It
- Enterprise AI should start from a map of actual tasks. Johnson says enterprises mostly adopt AI because of media hype or board pressure on the C-suite. WorkHelix instead surfaces jobs and tasks from an enterprise's raw data and maps bundles of tasks to generative-AI solutions. It then measures efficacy in a proper A/B test and reassesses. Clients learn, for example, that marketing holds far more manual work than they thought. The host cites a co-founder's claim that this could unlock a trillion-dollar opportunity.
- Behavioural friction explains the rush into AI rollups. Johnson says investors buy service-heavy companies and run them on AI-native systems because changing an existing enterprise's behaviour is so hard. AI startups' professional-services line items reflect the same friction. One CEO told him the firm had to mandate AI use: "every Friday is AI day."
GeoHussarCan Iran defend itself and retaliate effectively? Probably not. #iran #tehran #defense #geopolitics
- Iran's air defences were already gutted before this strike. He says over 200 aircraft entered Iranian airspace with none shot down, only a couple of drones reportedly downed. Iran relies on Russian S-300s, most of which were destroyed in April and October 2024 along with solid-fuel production equipment, and the system struggles to track the stealthy F-35s used.
- Infiltration and lost leadership leave Iran unable to retaliate effectively. He says top nuclear scientists and IRGC commanders were killed easily, and a drone base set up inside Iran targeted its missile infrastructure for when it tried to retaliate. Iran's only response so far was about 100 drones.