Ken Tabor

Open source AI is thriving even in a world of closed foundational LLM success. WHAT'S GOING ON Hugging Face’s Spring 2026 report is useful because the authors show where a lot of open AI work is happening. It’s very interesting to me that HF is a place where practical development lives. Tech isn’t always about inventing new, heavy-metal, frontier models. HF is a vibrant community where datasets, experiments, and specific models show up first. KEY TAKEAWAYS * Real value is in fine-tunes, optimizations, and evals. * The open model landscape is HUGE. Learn to filter and navigate it. * Smaller models matter are crucial in the reality of constrained hardware, latency KPIs, and costs. * U.S. AI R&D labs are releasing powerful open LLMs like OpenAI gpt-oss, Google Gemma, and NVidia Nemotron. * Other countries will adopt sovereign intelligence. * We need to build application architectures that can integrate, swap or route, guard rail, and monitor models over time. WHY THIS MATTERS AI Engineering is becoming a real job. Professional discipline remains a valuable differentiator. Picking a good model is only part of the battle. Practice with open models to gain intuition for creating evals, fine-tuning, optimizing, and knowing when to migrate between them. Read the article because it gives a clearer picture of where the builder energy in open AI is moving. WANT MORE? The article is “State of Open Source on Hugging Face: Spring 2026.” For my part, I’ve described Hugging Face of a cross between “Github + Stack Overflow.” Take time to visit the site and learn your way around it. If you find other good AI or engineering reads, send them my way! huggingface.co/blog/huggingfa…

Technical challenges like these have been a huge win for companies in the past. Especially for capability awareness of the company side, and competence of potential hires on the other side. Check it out if you want to prove your mettle!

Does AI feel suddenly everywhere, and hard to relate to? Here’s one way of thinking about how we got here. I put this “Brief History of Generative AI” slide together as a simple orientation point. Maybe you've felt how quickly AI moved from something speculative to something that is now sitting much closer to everyday life, especially for product managers, software engineers, and UX designers. Maybe you've been wondering where the foundational ideas behind Generative AI came from. Note: this slide is not meant to be a definitive history. It is a short sequence of major milestones that, from my point of view, helps explain how we got here. AI feels different this time, and gaining some shared context is a worthy goal. There is a lot of uncertainty at work these days. I get it. Now is a time to respond with curiosity, experimentation, and adaptability. Leverage AI to be more awesome!

Dude - I know you suffer for your art. Here's a great idea - everyone please leverage the detailed DD as grounding knowledge for AI analysis. Anyone who embraces "Chat with the PowerPoint" is going to have fantastic success in the Age of AI. Our KookBot solves this! Anyone can launch the Custom GPT, copy in the bear case text, and ask AI: "I’ve read a negative post related to AST SpaceMobile. Glad to read your analysis of it considering where it might be correct, and where it might be missing the point." x.com/KenTabor/statu…

A great artist must suffer the most blistering critiques. $ASTS x.com/thekookreport/…

$ASTS I have to say this. If you’re selling or short AST Spacemobile here, you have to be the dumbest fucking investor or most retarded trader to ever exist. If you have a fireplace in your house/rental, you should literally just light the MF’er and throw all your money in now.

@spacanpanman It'll be a massive catalyst. Reducing risk, removing uncertainty, unlocking revenue potential, updating every investor AST is a viable company. Sadly, Ronaldo says otherwise 🤣

$ASTS: When the FCC approves AST SpaceMobile for full US commercial service, what will be the one-day stock price reaction?

Maybe you've heard about edge-compute AI. It's very interesting to think of running a modern open-source LLM on a handheld device. As an alternative to a pay-to-play cloud provider. OSS LLMs include Mistral, gpt-oss, Nemotron, Gemma, and Qwen. That's just text, I'll be curious to see GenMedia too. Such a device might enable your personal R&D as you code AI backed features in your agentic application. The team at @TiinyAILab are building such a machine. I've just backed them through Kickstarter, and super curious to get my hands on this kit! kickstarter.com/projects/tiiny…

@Chartradamus Found and read these books over past few years. Your post made me think - I need to read these again soon! Have fun - enjoy!

Recently got back into reading - moreso the genres of science fiction and fantasy. Curious. Any book recommendations?

@KrakTrades LOL. True. We’ve seen things in the SPAC world tho.

I like $ASTS and I’m long. But man, this community is so emotional and irrational sometimes.

@ASTS_Investors How much fun is this? So much fun seeing management execute on the plan.

$ASTS

You'll want to read this book if you're curious about leveraging AI agents in the workplace. "You to the Power of Two" by Bradley and Tapscott was published December 2025. It's main thesis is formed around the idea that "Perhaps you're excited by the near-term prospect of personal AI that augments your life. In a few years, these digital extensions of you will be infinitely more capable, with perfect memory and near-limitless intelligence." (page 3) I've pulled a few quotes from the book to get you interested in it: "As we delegate tasks and decisions to AI, we must first define the values and the principles that guide our lives. What information should we retain in our memory verses offloading to our AI agent?" (page 43) "In an AI-driven economy, humans are developing what we describe as 'AI superpowers' that fundamentally enhance our capabilities at work. Say you were never very good with numbers but suddenly gained the ability to perform instant data analysis." (page 81) "...employees may argue that an AI copilot is akin to a personal assistant tailored to their working style and developed through years of interaction. Losing that AI upon leaving an employer could feel akin to losing an essential part of their professional tool kit." (page 107) "If your organization isn't already reskilling and upskilling employees to work alongside AI systems, and if you're not cultivating trust in the teams responsible for executing AI-supported strategies, then you're already a laggard." (page 300) Bradley and Tapscott cover a wide range of issues about AI extending past work life. It's the main point given this site, but you'll find it's a sweeping survey. Glad to have read this book. It's clear, well organized, and an easy read in the evening. Go buy a copy if you're curious to dig in! Let me know what you think in the comments below.

Wow. Amazing numbers and big money. I started coding with Codex last Fall. It’s so much fun! Fun like when I was a boy first starting with Basic, DOS, Edlin, on the IBM PC. “Weekly Codex users have more than tripled since the start of the year to 1.6M.”

@endless_frank With a hilariously unexpected social proof-point from @spacanpanman himself! 😉

$ASTS Zerohedge with a piece about Starlink. Just regurgitated stuff from yesterday’s conference. Currently running 4mbps of speed with a goal of getting to 150mbps by launching over 14,000 more satellites for D2C services. We knew Starlink would be a D2C player so carry on.

@oliwalkerjones Amazing machine - can't wait to see Joby flying around nearby DFW Intl! Did yall fly near the Suzuka track? Got me excited for F1 season opener.

This reel posted by Toyota over on Instagram is a thing of beauty! $JOBY Shot during our Mount Fuji demo at the end of last year

@blueorigin @AST_SpaceMobile We love seeing the progress !

Look what rolled in this morning! The payload fairing for New Glenn's third launch just arrived at the Integration Facility at LC-36 with @AST_SpaceMobile's BlueBird satellite safely secured inside.

@tottaway22 @AST_SpaceMobile Impressive seeing this. All the work that’s planned, and executed on, is extraordinary!

Live now!!!! The Blue Origin New Glenn Flight 3 Faring with the @AST_SpaceMobile $ASTS BlueBird 7 inside headers toward LC-36 to wait for booster. Launching soon!

@CatSE___ApeX___ Awesome! I’ve never seen this process before.

$ASTS Watching the fairing rolling out live.

Could be an amazing act of service for some!

@spacanpanman Sofer talked about beginning Beta service soon. Real user data is going to hit their dashboards soon. Loads of experiments to test and learn on!

$ASTS: 🚨 Great interview of Roy Sofer, SVP of Engineering and head of the R&D office in Israel for AST SpaceMobile. Definitely take the time to read entire interview! Summary: ✅ Overview & Company Scale (Israel R&D center, vertical integration) ✅ Why D2D Now (what changed since 1990s satellite phones) ✅ Key Tech Breakthrough (standards compliance, Doppler, huge phased array) ✅ Performance & Interference (broadband user experience, beamforming, competition) ✅ Market Opportunity (coverage gaps, willingness to pay, TAM claim) ✅ Proof + Funding Path (early demos, investors, BlueWalker/BlueBird build-out) ✅ Israel Execution Model (in-house hardware, supply chain, “space factory”) ✅ Roadmap & Future Uses (launch cadence, coverage timeline, defense/IoT/radar) Original podcast here: open.spotify.com/episode/2mgazN…