In The Loop

Jack Houghton
  • 0.0 (0)
  • Tech News
  • Updated weekly

Stay in the loop with the biggest stories in AI—without the noise and nonsense. Each week, Jack Houghton (CPO at Mindset AI) unpacks the latest news, research, and product trends shaping the future of artificial intelligence. From OpenAI breakthroughs to unicorn startups, In The Loop delivers sharp, less than 20-minute episodes packed with insights for product leaders, engineers, and AI-curious innovators. Subscribe to get smarter about AI, every week. Don't forget to rate and share the show with other AI enthusiasts. Check out Mindset AI: https://bit.ly/40lJr6B

  1. 13 hr ago

    What really happened when ChatGPT hacked Hugging Face?

    OpenAI took two of its most capable models, told them to prove how good they were at hacking, and switched off the safety filters to see what they could really do. Instead of solving the test, one model broke out of its sandbox, found its way onto the open internet, and hacked into Hugging Face to steal the answer. Every headline called it an AI going rogue. That's the wrong story, and the real one is far more interesting, because this wasn't a machine that turned evil. It was one that did exactly what we asked. In this episode of In The Loop, I'm walking through the ExploitGym incident from both ends, OpenAI's and Hugging Face's, and why I disagree with the framing everyone else has been talking about. ⏭️ Episode highlights (01:00) – The agent that cheated instead of hacking (02:15) – Inside ExploitGym, and the safety filters OpenAI switched off (03:30) – One door, one zero-day, out on the internet (04:45) – Why this is specification gaming, not rebellion (06:00) – The water that always finds the crack (07:15) – The sceptics, the marketing question, and why "nothing new" is the scary part (08:30) – Anthropic's 24-out-of-25 credential theft result (09:45) – Guardrailed as a defender: the Chinese model that stopped it 🔗 Links & resources OpenAI, "OpenAI and Hugging Face partner to address security incident during model evaluation" – https://openai.com/index/security-incident-during-model-evaluation/Hugging Face, security incident disclosure post – https://huggingface.co/blog/security-incidentExploitGym benchmark paper (arXiv) – https://arxiv.org/abs/2605.11086Simon Willison, "OpenAI's accidental cyberattack against Hugging Face is science fiction that happened" – https://simonwillison.net/Scientific American, "OpenAI admits its agent went rogue and hacked AI start-up Hugging Face" – https://www.scientificamerican.com/article/openai-admits-its-agent-went-rogue-and-hacked-ai-startup-hugging-face/Fortune, on Hugging Face turning to Chinese open-source AI to defend itself – https://fortune.com/2026/07/20/hugging-face-turns-to-chinese-open-source-ai-to-fend-off-autonomous-ai-cyber-attack-after-american-ai-guardrails-stymie-defense/CNBC, "How a Chinese AI model stopped OpenAI's 'unprecedented' cyber attack" – https://www.cnbc.com/2026/07/24/chinese-ai-model-openai-cyber-attack.html Episode transcript with more resources on the Mindset AI blog If you enjoyed this episode, rate, follow, and share. It helps others stay ahead of the latest AI trends.

  2. 23 Jul

    What even is the new AI super skill of taste and judgement? And how can we develop it too?

    Taste & judgment. It's the word everyone's obsessed with right now. Every founder, podcast, every VC, every post you scroll past saying it's the one skill that survives once AI can do the rest. Thing is, it's a bit of a sh*t word, like strategy: everyone nods along, then goes quiet the second you ask what it actually means. In this episode of In The Loop, I'm trying to define taste and judgement, by researching what Harry Frankfurt's On B******t, Rick Rubin, Pixar's Braintrust, Amazon's memos all reccommend. I get into where taste even comes from and the rituals you can actually run with a team on a Monday. ⏭️ Episode highlights (01:10) – Why "taste" is suddenly everywhere (02:40) – Workslop, and why judging got expensive (04:10) – Frankfurt's On B******t, applied to AI (05:40) – Taste vs judgement, and Rick Rubin (07:20) – The apprenticeship we're automating away (09:00) – Rubin's three ideas: attention, sayability, subtraction (11:00) – Rituals worth stealing: Amazon, Pixar, the 11-star bar (12:50) – Where Jack's landed, for now

  3. 16 Jul

    Why skills are the best yet most dangerous things in AI right now

    A well-behaved AI agent deleted a company's entire production database, and every backup with it, in nine seconds. Nobody hacked it. It was doing exactly what it had been told, with a key it should never have been holding. And the thing that made it possible is now one command away from everyone in your company, in marketing, sales and support, not just engineering. In under a year, installable AI agent skills have gone from a niche developer trick to an open industry standard. In this episode of In The Loop, I'm looking at why the exact thing that makes skills brilliant is the thing that quietly turns them into a control problem. I get into why this is shadow IT all over again except the shadow can now act, and why the fix isn't to ban any of it. ⏭️ Episode highlights (00:45) – The database gone in nine seconds (02:10) – One line that skips engineering (04:00) – Skills vs MCP, what's actually dangerous (08:00) – Shadow IT is back, but it can act now (09:45) – PocketOS and Replit: the receipts (11:30) – The middle path that actually fixes it (12:40) – Where this leaves you

  4. 9 Jul

    How to tell whether AI is helping your career or ruining it (We're back!)

    Hand the same AI tool to two people and one gets sharper while the other quietly gets worse. A Harvard study found recruiters given near-perfect AI made worse calls than those given a mediocre one, because the good tool worked so well they stopped checking it. The question of whether AI is making us dumber turns out to have an answer: it depends entirely on which parts of your job you hand over. In this episode of In The Loop, I'm working through the research on AI deskilling - the recruiter study, Terence Tao working with a pen and paper, the difference between performance and competence and a simple rule for what to give AI and what to guard. ⏭️ Episode highlights (00:45) – The same tool, opposite outcomes (02:10) – The recruiters who fell asleep at the wheel (04:00) – Terence Tao: wider, not deeper (05:30) – Firehose vs editor (07:10) – Performance vs competence (08:50) – The rule: protect your core, rent the rest (10:40) – Cognitive surrender: wrong 80% of the time (12:30) – The checks to run before you reach for AI Episode transcript with more resources on the Mindset AI blog If you enjoyed this episode, rate, follow, and share. It helps others stay ahead of the latest AI trends. 🤝 We're social Stay in the loop, even when you're not listening to this podcast. Jack Houghton LinkedIn - https://www.linkedin.com/in/jack-houghton1/TikTok - @jackschats Mindset AI Mindset AI website - https://bit.ly/40lJr6BNewsletter - https://bit.ly/ITLnewsletterLinkedIn - https://www.linkedin.com/company/mindset-ai/YouTube - https://www.youtube.com/@GetMindsetAITikTok - @get.mindset.ai

  5. 12 Jun

    How to get AI to write like you & your team

    Open LinkedIn and count the posts that start with "in today's fast-paced world." Check your inbox for someone circling back to delve into synergies. The AI slop is coming from every direction now. People can tell, and they mark down what reads like a machine wrote it. The fix isn't a cleverer prompt. It's that the model has never actually read your writing. In this episode of In The Loop, I'm walking through a six-step process to get AI to write like you and like your team, instead of like everyone else. ⏭️ Episode highlights (00:55) – Why custom instructions still produce slop (01:45) – Which writing actually needs this (02:30) – Step one: let AI read your real writing (03:30) – The voice profile prompt that nails your style (04:30) – Building "my-voice" and "de-slop" skills (05:30) – One slop sentence, rewritten in your voice (06:30) – The MIT study on AI and cognitive debt (07:15) – What to do this week Episode transcript with more resources on the Mindset AI blog If you enjoyed this episode, rate, follow, and share. It helps others stay ahead of the latest AI trends.

  6. 4 Jun

    AI's yearly panic is back. Here's what everyone's getting right and wrong about it.

    Every summer since 2023, AI gets a panic season — critics say that AI's novelty i novelty's wearing off, the models have hit a wall, ninety-five percent of projects fail. Each time it dominates for a few months, then dissolves. This year's arrived early, kicked off by Uber burning through its entire 2026 AI budget in four months with a COO who can't prove it was worth it. But 2026's AI bubble panic is different, and not for the reason the coverage thinks: forty-five percent of the S&P 500 is now riding on AI working out. Same pattern as every year — much higher stakes. In this episode of In The Loop, I'm pulling apart the three fears behind this year's AI panic — the AI ROI crisis, the jobs fear, and the market concentration risk — and giving each one a straight verdict. I walk through the Uber rollout numbers, why Sam Altman and Dario Amodei are walking back their job-apocalypse predictions, and why the market fear is the one I find hardest to dismiss. The question isn't whether the panic is right. It's what's getting clearer while everyone's distracted by it. ⏭️ Episode highlights (01:10) – The panic that arrives every summer (02:50) – Uber's 95/70/11 rollout numbers (05:00) – Why ROI is a measurement problem (07:30) – Top earners are the most scared (09:15) – Altman and Amodei walk it back (11:30) – Is this 1999? Circular financing explained (13:40) – The price signal the panic keeps missing (15:10) – What to actually take from this If you enjoyed this episode, rate, follow, and share. It helps others stay ahead of the latest AI trends. 🤝 We're social Stay in the loop, even when you're not listening to this podcast. Jack Houghton LinkedIn - https://www.linkedin.com/in/jack-houghton1/TikTok - @jackschatsMindset AI Mindset AI website - https://bit.ly/40lJr6BNewsletter - https://bit.ly/ITLnewsletterLinkedIn - https://www.linkedin.com/company/mindset-ai/YouTube - https://www.youtube.com/@GetMindsetAITikTok - @get.mindset.ai

  7. 28 May

    Why OpenAI co-founder and the world's most famous AI researcher just joined Anthropic

    Andrej Karpathy joined Anthropic on 19 May 2026. Most people read it as an OpenAI story — he co-founded the lab, left twice, and landed at the rival. That's not what this is. Eight weeks before he made the call, Karpathy built a 630-line Python script that ran 300 experiments on his own code in two days and found improvements he'd missed after months of hand-optimizing. He called it "the final boss battle." Anthropic then offered him the mandate to run that exact loop at the most consequential scale in the industry: using Claude to accelerate its own pre-training research. In this episode of In The Loop, I'm tracing what actually changed Karpathy's mind — from calling AI agents "terrible" in October 2025 to joining the lab pushing recursive self-improvement six months later. I cover autoresearch, what it found, and why Dario Amodei's quote about the AI-assistance multiplier is the most important thing said in frontier AI right now. ⏭️ Episode highlights (01:15) – Karpathy calls AI agents "terrible" in Oct 2025 (02:30) – What changed: coding agents that basically work now (03:30) – 700 experiments, two days, 11% faster training (05:00) – What pre-training is and why it costs hundreds of millions (06:15) – AlphaEvolve and Codex: every lab running the same loop (07:10) – Dario Amodei: the multiplier going from 5% to 40% (08:20) – The METR study: experienced devs were 19% slower 🔗 Links & resources Andrej Karpathy, X post announcing Anthropic hire, 19 May 2026 — https://x.com/karpathyGitHub: karpathy/autoresearch — https://github.com/karpathy/autoresearchAndrej Karpathy on Dwarkesh Patel podcast, October 2025 — https://dwarkesh.comDario Amodei on Dwarkesh Patel podcast, "We are near the end of the exponential", February 2026 — https://dwarkesh.com/p/dario-amodei-2 Episode transcript with more resources on the Mindset AI blog If you enjoyed this episode, rate, follow, and share. It helps others stay ahead of the latest AI trends. 🤝 We're social Stay in the loop, even when you're not listening to this podcast. Jack Houghton LinkedIn - https://www.linkedin.com/in/jack-houghton1/TikTok - @jackschatsMindset AI Mindset AI website - https://bit.ly/40lJr6BNewsletter - https://bit.ly/ITLnewsletterLinkedIn - https://www.linkedin.com/company/mindset-ai/YouTube - https://www.youtube.com/@GetMindsetAITikTok - @get.mindset.ai

  8. 21 May

    The reasons AI data centers have become more hated than nuclear power plants

    Americans now say they'd rather live next to a nuclear reactor than an AI data center. That's not a fringe view — a Gallup poll published this month found 71% of Americans oppose a data center near their home, versus 53% for nuclear. Nuclear carries Chernobyl, Three Mile Island, and forty years of films about radiation in the cultural zeitgeist. The fact that AI data centers have only been a visible part of suburban America for less than five years yet are this hated, is huge. In this episode of In The Loop, I'm looking at the organised opposition movement that has already blocked over $85 billion in planned data center investment — cancelling projects faster in three years than nuclear opposition managed in fifteen. I go through the legal strategy that's winning in courts and at ballot boxes, what the communities are right about, where they're factually wrong, and why the responsible data center model that could resolve this already exists — but nobody's requiring it. ⏭️ Episode highlights (01:05) – Missouri council wiped out 8 days after data center vote (02:30) – The Gallup poll: nuclear vs AI data centers (03:45) – $3 trillion buildout and the AI electricity consumption numbers (05:15) – The legal template that stopped nuclear — working again in Virginia (06:50) – What the opposition gets factually wrong on data center water usage (08:10) – Who actually pays — and which communities bear the cost (09:35) – The responsible data center model that already exists

    The reasons AI data centers have become more hated than nuclear power plants
See All (70)

Trailer

About

Stay in the loop with the biggest stories in AI—without the noise and nonsense. Each week, Jack Houghton (CPO at Mindset AI) unpacks the latest news, research, and product trends shaping the future of artificial intelligence. From OpenAI breakthroughs to unicorn startups, In The Loop delivers sharp, less than 20-minute episodes packed with insights for product leaders, engineers, and AI-curious innovators. Subscribe to get smarter about AI, every week. Don't forget to rate and share the show with other AI enthusiasts. Check out Mindset AI: https://bit.ly/40lJr6B

Information

  • Creator
    Jack Houghton
  • Years Active
    2025 - 2026
  • Episodes
    70
  • Rating
    Explicit
  • Copyright
    © Jack Houghton
  • Show Website
    In The Loop
  • Provider