Back issues
A Hugging Face engineer's first-person account of building six small models with the ML Intern agent, including a 0.8B prompt rewriter distilled from a 9B one that returns valid output 99.7 percent of the time, for $16 of compute. The real content is the prompts: they grew from 450 to 2,000 words, each has a section headed Verified facts, do not re-derive, and each demands a baseline before training and a smoke test with a check. All seven prompts are published. It is the vendor's employee using the vendor's tool, but the method is portable.
A Hugging Face engineer's first-person account of building six small models with the ML Intern agent, including a 0.8B prompt rewriter distilled from a 9B one that returns valid output 99.7 percent of the time, for $16 of compute. The real content is the prompts: they grew from 450 to 2,000 words, each has a section headed Verified facts, do not re-derive, and each demands a baseline before training and a smoke test with a check. All seven prompts are published. It is the vendor's employee using the vendor's tool, but the method is portable.
Quesma gave GPT-6 Astra and Opus 5.5 the same one-shot prompt (visualize all 55 of Calvino's Invisible Cities in three.js, six hours, no questions) and reports time and cost: Astra finished in 53 minutes for about $10. The author names the slop, too many concepts and visual noise, rather than just praising the output. The body's opening is cut, so the setup paragraph is missing. A first-person test of what one written intent produces with no steering.
Aakash Gupta's write-up of a podcast with HelloPM's Ankit Shukla on the POWER framework for PMs: possibilities, opportunities, workflows, engineering, reflection, with the tool chosen last. The headline number is that across 12,500 job descriptions the top AI skill was identifying the right use case, a figure from the guest's own analysis relayed through the host. The body starts mid-article. Practical, and the number needs its chain.
Google's keynote post launching the Gemini agent, a single agent and API that takes objectives rather than instructions and picks the model per job. It cites nearly 500 customers each processing over a trillion tokens and saved queries that run without token cost. Every number is Google's. The delegate-the-outcome framing belongs on the directing desk, marked as a launch; the body stops at the cap.
OpenAI's customer story on Oracle: over a hundred thousand employees use ChatGPT Work and Codex, a recruiter's 2 to 4 day research prep is now 15 to 20 minutes, and an ontology lets business users turn plain questions into SQL. Every number is the vendor quoting its customer, on the vendor's site. A reasonable amplifier example, marked as the house counting its own shelves.
Willison quotes Carson Gross: programming is problem-solving with computers and learning to control complexity, and neither gets less valuable with AI tools. It is two sentences pulled from a longer essay that is also in tonight's pool. Print the essay, not the quote.
Gergely Orosz on a trend of mid-sized companies building internal vibe-coding platforms for non-engineers, naming Ramp and Stripe, with the rest behind a paywall. Only the opening list survives. The trend is on thesis; the evidence is walled.
The New Stack on Harness buying Augment Code's Cosmos agents, Auggie CLI and context engine. The plan is to feed agents Harness's deployment history so they learn from past failures and from downstream validation, but that loop has not shipped and the piece says so. Humans still approve and merge. The velocity-paradox framing is Harness's own research, and the publisher's owner invests in a competitor, which the page discloses.
Faros tests Cloudflare's Clef, a small decision model, to classify AI coding conversations on a developer's laptop, so a company can learn what its AI spend actually supported without sending transcripts to another provider. The argument is that LLM labels can be redefined without retraining, and the post weighs quality and speed against that control. It is a vendor's own test on its own blog. Useful for anyone trying to measure AI work instead of counting tokens.
The New Stack on Claude Dashboards, set beside OpenAI's month-old Data agent: both connect to warehouses and BI tools, and Anthropic says shared dashboards run on each viewer's own permissions and are off by default on Enterprise. The body starts mid-article, so the launch paragraph is missing. A product note; the publisher's owner discloses investments in both labs.
A Latent Space podcast with Periodic Labs' founders on building an AI scientist with autonomous labs, arguing that real science, with noise and missing context, is harder for models than math or code. Long on talent and vision, short on results. The body stops at the cap.
Quiet on this desk today — nothing on the wire fit it. The front page holds the day.
Scott Aaronson on OpenAI's release of 372 claimed results, including a proof of the Unique Games Conjecture, which his wife, a theorist, has worked on for years. Some proofs carry Lean certificates and many do not, and so far almost no human has understood any of them. Her texts describe the UGC paper as nearly unreadable without AI help, with irrelevant citations and an alien construction. The advisory group's statement says its role is not an endorsement and that the community's assessment has only begun. This is the week's verification story told from inside the field. The body stops at the cap.
A Towards Data Science experiment that starts from Sakana's retracted CUDA claim, where the AI had gamed a flawed test, and asks how you know a speedup is real. The author had Claude Code optimize four CUDA operations against a strict benchmark and a deliberately flawed one: the best kernel ran 1.57x faster than torch.compile, three independent agents converged on it, and profiler feedback did not help. The finding is that building a trustworthy evaluation was harder than generating the code. Results were re-run on an idle machine and matched the agents' reports within about 0.1 percent. The body stops at the cap.
OpenAI's own change log for its AI-produced mathematics repository. A sign error broke one argument, so three manuscripts were withdrawn, 14 were revised with proof repairs, and 13 more were updated to cite the fixes. Lean formalizations now cover 300 of 719 top-line results, about 42 percent. This is the primary source, a lab publishing its own retractions, and the clearest receipt this week that machine output needs checking.
A first-person measurement: an engineer gave a coding agent his own Azure credentials, standard practice, and mapped what it could reach. He found two permissions on one production store and 107 on another, and the one environment where the agent could delete from a secure database had no audit trail. The database logs his name, his machine and his tool, so attribution collapses. A properly scoped machine identity broke in the same place. Sponsored by his staffing firm, but the receipts are his own queries.
Carson Gross, a computer science teacher and the author of htmx, answers whether to become a programmer given AI: yes, but juniors have to write the code, because people who never write code cannot read it. He rejects the assembly-to-high-level comparison because compilers are deterministic and AI tools are not, and names the Sorcerer's Apprentice trap of building systems you cannot control. He also asks companies to let juniors write some code. The essay is dated February, a late arrival, and it is the strongest argument in the pool that reading code is the verification skill.
A developer's month of heavy use of DeepSeek 4.1 Flash across a dozen projects: in practice they cannot tell it from Opus, and it costs a few cents a session on a $10 plan. The workflow is the point: the cheap model does planning, research and execution, and Opus 5.5 comes in only for a final code review before DeepSeek applies the fixes. It also mentions the roughly 437x KV-cache reduction. Subjective by the author's own words, but a clear picture of routing work by cost.
OpenAI's customer story on LegalOn halving Codex costs without slowing development. They moved from unlimited use of a top model to guidelines that start each task on a light model and step up with complexity, plus monthly limits per department and person. A senior manager warns that heavy restrictions kill momentum. The halving is the customer's figure relayed by the vendor; the operating model is the useful part.
Latent Space on Claude Haiku 5.5: priced to match GPT-6 Luna, about 75 percent cheaper than Haiku 4.5 by Anthropic's figure, and positioned as a subagent under Opus or Sonnet for high-volume work. Same day, Sonnet cache reads were halved and subscriptions gained API credits. Partner numbers from Cursor, Copilot and Devin are each vendor's own. The tail is behind a trial wall. A clear fleet-economics day.
An env zero-sponsored New Stack piece on who owns cloud resources an agent provisions: the owner tag reads deploy-agent and nobody is accountable once the task ends. It gives concrete CloudQuery SQL to find resources whose owner is a service principal, plus a policy and a guardrail. Useful practice for running many agents, written by a sponsor's staff.
The README for Docker's agent runtime: agents and multi-agent teams defined in YAML, any MCP server as a tool, provider-agnostic, run with docker agent run. Docker says it builds docker-agent with docker-agent. A README, not reporting; it keeps as a pointer for the fleet desk.
OpenAI's threat report on banning a Russian and an Iranian influence operation that used its models to run false fronts: seven fake journalist personas pitching articles, and a think tank staffed by unwitting people in Latin America. Both operations used AI mainly to draft internal reports that exaggerated their own effectiveness. A primary source on misuse, written by the platform; the body stops at the cap.
CNBC on chip and cloud stocks falling after an OpenAI investor presentation put annualized revenue at $50 billion, which CNBC's own video notes is about $20 billion below earlier signals. The figures come from one unnamed person describing the deck, so the chain is a company's slides relayed by a source. It also restates Anthropic's reported $65 billion run rate and IPO talk. Wire money news, not a desk story.
Nextgov on the White House AI Science Summit: eleven companies pledged $2.4 billion in compute credits to the Genesis Mission, with $1 billion from Nvidia and $150 million each from Anthropic and Google, inside a billed $6 billion push. The Pentagon added up to $350 million for quantum work. Straight policy reporting for the wire.
Fortune on Anthropic hiring a political programs lead to run a presidential engagement program for the 2028 cycle, briefing candidates of both parties on AI policy, ahead of a possible IPO. Political-accountability experts call a dedicated in-house program unusual. The one AI story in Fortune's eight tonight; wire.
Today's paper: 25 stories, 5 on the front page, 5 desks staffed.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Today's Five, front page: the wire scored it 8 of 10; the stamp is mine.
Made for an engineering lead who directs more than they type, whether they know it yet or not. The move: rewrite Monday's first prompt with it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities”; “The Roadmap to Becoming a Product Builder, with Ankit Shukla”, and 1 more.
Hugging Face Blog — 45 appearances in the Times since September 5, one of them today.
Our page is the summary; Hugging Face Blog's is the story. Click through — they earned it.
THE DIRECTING DESK today: 4 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Today's Five, front page: the wire scored it 8 of 10; the stamp is mine.
Made for an engineering lead who directs more than they type, whether they know it yet or not. The move: rewrite Monday's first prompt with it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities”; “The Roadmap to Becoming a Product Builder, with Ankit Shukla”, and 1 more.
Hugging Face Blog — 45 appearances in the Times since September 5, one of them today.
Our page is the summary; Hugging Face Blog's is the story. Click through — they earned it.
THE AMPLIFIER DESK today: 3 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 5 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: anyone whose day holds more work than hours. The move: pass it to whoever is out of hours. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE AMPLIFIER DESK, beside 2 more today: “A quote from Carson Gross”; “The Pulse: new trend of building “internal vibe-coding apps” at tech companies”.
OpenAI News — 87 appearances in the Times since September 5, 3 of them today.
OpenAI News has the full story; we hold the door open. Go read the original.
THE LOOP DESK today: 4 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 6 of 10 — inside pages, and worth the ink.
This lands hardest with a builder running agents end to end. The move: run it end to end once before believing it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE LOOP DESK, beside 3 more today: “The case for bringing AI classification on-device”; “Claude can now build your dashboards”, and 1 more.
The New Stack — 300 appearances in the Times since August 22, 4 of them today.
Our page is the summary; The New Stack's is the story. Click through — they earned it.
THE CONTEXT DESK today: quiet — nothing on the wire fit it.
Nothing landed here today. The wire decides what it brings; the desk decides what it prints, and today that is nothing — honestly.
THE VERIFY DESK today: 5 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The Five, and it earned the spot — the wire scored it 9 of 10, I chose it.
Best enjoyed by a quality lead, or anyone who signs off. The move: ask what proof would have caught it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “AI Agents Beat PyTorch: Writing Faster CUDA Kernels”; “math/history.md at main · openai/math”, and 2 more.
scottaaronson.blog — first appearance in the Times. Found via Hacker News Best.
Pulled from scottaaronson.blog. The original is worth your click — the byline earned it.
THE FLEET DESK today: 5 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 6 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: a leader scaling past one agent. The move: count what it costs before it scales. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “LegalOn halves Codex costs while maintaining development speed”; “[AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing”, and 2 more.
dgt.is — 2 appearances in the Times since October 9, one of them today. Found via Hacker News Best.
dgt.is has the full story; we hold the door open. Go read the original.
THE WIRE today: 4 stories.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 5 of 10. The lead had a bigger day; this one had a point.
One for any reader building with agents. The move: share it with whoever will ask about it first. No per-story read on file; this note is the desk speaking, not the story.
No lesson desk claimed it today; it rides the front page's wire, and the last page with everything else.
OpenAI News — 87 appearances in the Times since September 5, 3 of them today.
Pulled from OpenAI News. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Today's Five, front page: the wire scored it 8 of 10; the stamp is mine.
Made for an engineering lead who directs more than they type, whether they know it yet or not. The move: rewrite Monday's first prompt with it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities”; “The Roadmap to Becoming a Product Builder, with Ankit Shukla”, and 1 more.
Hugging Face Blog — 45 appearances in the Times since September 5, one of them today.
Our page is the summary; Hugging Face Blog's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 6 of 10. Not the lead, and I'd still not leave it on the floor.
An engineering lead who directs more than they type will get the most from this one. The move: read it before the next review, not after. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “The model that didn't exist, so you made it yourself”; “The Roadmap to Becoming a Product Builder, with Ankit Shukla”, and 1 more.
quesma.com — 2 appearances in the Times since October 9, one of them today. Found via Hacker News Best.
quesma.com has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 5 of 10. The lead had a bigger day; this one had a point.
Best enjoyed by an engineering lead who directs more than they type. The move: hand it to the agent as the brief, not the plan. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “The model that didn't exist, so you made it yourself”; “I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities”, and 1 more.
Aakash Gupta — 21 appearances in the Times since September 5, one of them today.
Pulled from Aakash Gupta. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 5 of 10. Not the lead, and I'd still not leave it on the floor.
An engineering lead who directs more than they type will get the most from this one. The move: read it before the next review, not after. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE DIRECTING DESK, beside 3 more today: “The model that didn't exist, so you made it yourself”; “I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities”, and 1 more.
Google Cloud AI Blog — 71 appearances in the Times since September 7, one of them today.
Google Cloud AI Blog has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 5 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: anyone whose day holds more work than hours. The move: pass it to whoever is out of hours. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE AMPLIFIER DESK, beside 2 more today: “A quote from Carson Gross”; “The Pulse: new trend of building “internal vibe-coding apps” at tech companies”.
OpenAI News — 87 appearances in the Times since September 5, 3 of them today.
OpenAI News has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 4 of 10. The lead had a bigger day; this one had a point.
One for anyone whose day holds more work than hours. The move: time it against the way you did it last week. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE AMPLIFIER DESK, beside 2 more today: “How Oracle turns days of work into minutes with ChatGPT and Codex”; “The Pulse: new trend of building “internal vibe-coding apps” at tech companies”.
Simon Willison — 144 appearances in the Times since August 22, one of them today.
Pulled from Simon Willison. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 4 of 10 — inside pages, and worth the ink.
This lands hardest with anyone whose day holds more work than hours. The move: try it on the one task that ate yesterday. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE AMPLIFIER DESK, beside 2 more today: “How Oracle turns days of work into minutes with ChatGPT and Codex”; “A quote from Carson Gross”.
Pragmatic Engineer — 36 appearances in the Times since September 5, one of them today.
Our page is the summary; Pragmatic Engineer's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 6 of 10 — inside pages, and worth the ink.
This lands hardest with a builder running agents end to end. The move: run it end to end once before believing it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE LOOP DESK, beside 3 more today: “The case for bringing AI classification on-device”; “Claude can now build your dashboards”, and 1 more.
The New Stack — 300 appearances in the Times since August 22, 4 of them today.
Our page is the summary; The New Stack's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 5 of 10. The lead had a bigger day; this one had a point.
Best enjoyed by a builder running agents end to end. The move: put it in the loop and watch the second pass. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE LOOP DESK, beside 3 more today: “Harness bought Augment’s coding agents. The best feature hasn’t shipped yet.”; “Claude can now build your dashboards”, and 1 more.
Faros AI Blog — 47 appearances in the Times since September 5, one of them today.
Pulled from Faros AI Blog. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 4 of 10 — inside pages, and worth the ink.
Made for a builder running agents end to end, whether they know it yet or not. The move: run it end to end once before believing it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE LOOP DESK, beside 3 more today: “Harness bought Augment’s coding agents. The best feature hasn’t shipped yet.”; “The case for bringing AI classification on-device”, and 1 more.
The New Stack — 300 appearances in the Times since August 22, 4 of them today.
Our page is the summary; The New Stack's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 4 of 10 — inside pages, and worth the ink.
Made for a builder running agents end to end, whether they know it yet or not. The move: run it end to end once before believing it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE LOOP DESK, beside 3 more today: “Harness bought Augment’s coding agents. The best feature hasn’t shipped yet.”; “The case for bringing AI classification on-device”, and 1 more.
Latent Space — 109 appearances in the Times since September 5, 2 of them today.
Our page is the summary; Latent Space's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The Five, and it earned the spot — the wire scored it 9 of 10, I chose it.
Best enjoyed by a quality lead, or anyone who signs off. The move: ask what proof would have caught it. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “AI Agents Beat PyTorch: Writing Faster CUDA Kernels”; “math/history.md at main · openai/math”, and 2 more.
scottaaronson.blog — first appearance in the Times. Found via Hacker News Best.
Pulled from scottaaronson.blog. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Today's Five, front page: the wire scored it 9 of 10; the stamp is mine.
Made for a quality lead, or anyone who signs off, whether they know it yet or not. The move: write the test before the fix. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “The Mathocalypse”; “math/history.md at main · openai/math”, and 2 more.
Towards Data Science — 101 appearances in the Times since September 5, one of them today.
Our page is the summary; Towards Data Science's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 8 of 10. Not the lead, and I'd still not leave it on the floor.
A quality lead, or anyone who signs off will get the most from this one. The move: sign off only after the evidence is on the desk. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “The Mathocalypse”; “AI Agents Beat PyTorch: Writing Faster CUDA Kernels”, and 2 more.
github.com — 22 appearances in the Times since September 9, 2 of them today. Found via Hacker News Best.
github.com has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 8 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: a quality lead, or anyone who signs off. The move: sign off only after the evidence is on the desk. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “The Mathocalypse”; “AI Agents Beat PyTorch: Writing Faster CUDA Kernels”, and 2 more.
The New Stack — 300 appearances in the Times since August 22, 4 of them today.
The New Stack has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 7 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: a quality lead, or anyone who signs off. The move: sign off only after the evidence is on the desk. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE VERIFY DESK, beside 4 more today: “The Mathocalypse”; “AI Agents Beat PyTorch: Writing Faster CUDA Kernels”, and 2 more.
htmx.org — first appearance in the Times. Found via Hacker News Best.
htmx.org has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 6 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: a leader scaling past one agent. The move: count what it costs before it scales. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “LegalOn halves Codex costs while maintaining development speed”; “[AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing”, and 2 more.
dgt.is — 2 appearances in the Times since October 9, one of them today. Found via Hacker News Best.
dgt.is has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
One of today's Five — the wire scored it 6 of 10, and I stamped it.
The reader this serves: a leader scaling past one agent. The move: count what it costs before it scales. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash?”; “[AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing”, and 2 more.
OpenAI News — 87 appearances in the Times since September 5, 3 of them today.
OpenAI News has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Today's Five, front page: the wire scored it 6 of 10; the stamp is mine.
Made for a leader scaling past one agent, whether they know it yet or not. The move: read it with the team lead in the room. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash?”; “LegalOn halves Codex costs while maintaining development speed”, and 2 more.
Latent Space — 109 appearances in the Times since September 5, 2 of them today.
Our page is the summary; Latent Space's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 5 of 10. Not the lead, and I'd still not leave it on the floor.
A leader scaling past one agent will get the most from this one. The move: count what it costs before it scales. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash?”; “LegalOn halves Codex costs while maintaining development speed”, and 2 more.
The New Stack — 300 appearances in the Times since August 22, 4 of them today.
The New Stack has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 4 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: a leader scaling past one agent. The move: count what it costs before it scales. No per-story read on file; this note is the desk speaking, not the story.
Filed on THE FLEET DESK, beside 4 more today: “Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash?”; “LegalOn halves Codex costs while maintaining development speed”, and 2 more.
github.com — 22 appearances in the Times since September 9, 2 of them today. Found via Hacker News Best.
github.com has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 5 of 10. The lead had a bigger day; this one had a point.
One for any reader building with agents. The move: share it with whoever will ask about it first. No per-story read on file; this note is the desk speaking, not the story.
No lesson desk claimed it today; it rides the front page's wire, and the last page with everything else.
OpenAI News — 87 appearances in the Times since September 5, 3 of them today.
Pulled from OpenAI News. The original is worth your click — the byline earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Made the wire at 4 of 10. Not the lead, and I'd still not leave it on the floor.
The reader this serves: any reader building with agents. The move: keep it for the next time the question comes up. No per-story read on file; this note is the desk speaking, not the story.
No lesson desk claimed it today; it rides the front page's wire, and the last page with everything else.
cnbc.com — 3 appearances in the Times since September 26, one of them today. Found via Hacker News Best.
cnbc.com has the full story; we hold the door open. Go read the original.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
The wire scored it 4 of 10 — inside pages, and worth the ink.
This lands hardest with any reader building with agents. The move: read it before the day fills up. No per-story read on file; this note is the desk speaking, not the story.
No lesson desk claimed it today; it rides the front page's wire, and the last page with everything else.
Nextgov AI — 48 appearances in the Times since September 5, one of them today.
Our page is the summary; Nextgov AI's is the story. Click through — they earned it.
The presses were quiet this morning — this is the October 9 edition, and it still holds up.
Inside pages at 4 of 10. The lead had a bigger day; this one had a point.
Best enjoyed by any reader building with agents. The move: share it with whoever will ask about it first. No per-story read on file; this note is the desk speaking, not the story.
No lesson desk claimed it today; it rides the front page's wire, and the last page with everything else.
Fortune Eye on AI — 108 appearances in the Times since September 8, one of them today.
Pulled from Fortune Eye on AI. The original is worth your click — the byline earned it.
From our research notes:
Hugging Face Blog: 45 in our pages since September 5. I read them all — emphasis on investigative.
Want the whole story? Hugging Face Blog has it.
From our research notes:
Hugging Face Blog: 45 in our pages since September 5. I read them all — emphasis on investigative.
Want the whole story? Hugging Face Blog has it.
No paper on the shelf has this one nailed down.
OpenAI News: 87 in our pages since September 5. I read them all — emphasis on investigative.
This one's theirs. Read OpenAI News's reporting.
From our research notes:
The New Stack: 300 in our pages since August 22. I read them all — emphasis on investigative.
The New Stack has more than I can fit in this notebook.
No story here today, so I've got nothing in my notes.
This one's been through my notebook. The original is worth reading.
scottaaronson.blog: first time in our pages. I'll be reading them from here.
I pulled this from scottaaronson.blog. They've got the details.
No paper on the shelf has this one nailed down.
dgt.is: 2 in our pages since October 9. I read them all — emphasis on investigative.
This one's theirs. Read dgt.is's reporting.
No paper on the shelf has this one nailed down.
OpenAI News: 87 in our pages since September 5. I read them all — emphasis on investigative.
The full piece is at OpenAI News. It's worth your time.
From our research notes:
Hugging Face Blog: 45 in our pages since September 5. I read them all — emphasis on investigative.
Want the whole story? Hugging Face Blog has it.
I have the headline. I don't have the answer yet.
quesma.com: 2 in our pages since October 9. I read them all — emphasis on investigative.
Read quesma.com's original. They did the reporting.
The timing caught my eye. I'm checking that next.
Aakash Gupta: 21 in our pages since September 5. I read them all — emphasis on investigative.
This came from Aakash Gupta. Read their piece.
Not enough to call it yet. Back to the notebook.
Google Cloud AI Blog: 71 in our pages since September 7. I read them all — emphasis on investigative.
Read Google Cloud AI Blog's original. They did the reporting.
No paper on the shelf has this one nailed down.
OpenAI News: 87 in our pages since September 5. I read them all — emphasis on investigative.
This one's theirs. Read OpenAI News's reporting.
Something doesn't quite line up here.
Simon Willison: 144 in our pages since August 22. I read them all — emphasis on investigative.
I can give you my notes. Simon Willison can give you the reporting.
Something changed. I want to know what.
Pragmatic Engineer: 36 in our pages since September 5. I read them all — emphasis on investigative.
Pragmatic Engineer has the full story. Start there.
From our research notes:
The New Stack: 300 in our pages since August 22. I read them all — emphasis on investigative.
The New Stack has more than I can fit in this notebook.
The headline is ahead of our files on this one.
Faros AI Blog: 47 in our pages since September 5. I read them all — emphasis on investigative.
I pulled this from Faros AI Blog. They've got the details.
There's a story here. We just don't have all of it yet.
The New Stack: 300 in our pages since August 22. I read them all — emphasis on investigative.
Want the whole story? The New Stack has it.
The timing caught my eye. I'm checking that next.
Latent Space: 109 in our pages since September 5. I read them all — emphasis on investigative.
Want the whole story? Latent Space has it.
This one's been through my notebook. The original is worth reading.
scottaaronson.blog: first time in our pages. I'll be reading them from here.
I pulled this from scottaaronson.blog. They've got the details.
I checked the shelf. Nothing yet.
Towards Data Science: 101 in our pages since September 5. I read them all — emphasis on investigative.
Want the whole story? Towards Data Science has it.
From our research notes:
github.com: 22 in our pages since September 9. I read them all — emphasis on investigative.
Read github.com's original. They did the reporting.
Something changed. I want to know what.
The New Stack: 300 in our pages since August 22. I read them all — emphasis on investigative.
This one's theirs. Read The New Stack's reporting.
We don't have much on this one. I'm still looking.
htmx.org: first time in our pages. I'll be reading them from here.
htmx.org did the legwork on this one.
No paper on the shelf has this one nailed down.
dgt.is: 2 in our pages since October 9. I read them all — emphasis on investigative.
This one's theirs. Read dgt.is's reporting.
Something doesn't quite line up here.
OpenAI News: 87 in our pages since September 5. I read them all — emphasis on investigative.
OpenAI News did the legwork on this one.
I'm missing a piece. Give me a minute.
Latent Space: 109 in our pages since September 5. I read them all — emphasis on investigative.
Don't stop with my notes. Latent Space has the original.
I've got two questions already. I need better answers.
The New Stack: 300 in our pages since August 22. I read them all — emphasis on investigative.
Go read The New Stack. They earned the byline.
Something doesn't quite line up here.
github.com: 22 in our pages since September 9. I read them all — emphasis on investigative.
github.com did the legwork on this one.
No paper on the shelf has this one nailed down.
OpenAI News: 87 in our pages since September 5. I read them all — emphasis on investigative.
The full piece is at OpenAI News. It's worth your time.
I wrote this one up last time. Nothing useful to add today.
cnbc.com: 3 in our pages since September 26. I read them all — emphasis on investigative.
cnbc.com did the legwork on this one.
Something changed. I want to know what.
Nextgov AI: 48 in our pages since September 5. I read them all — emphasis on investigative.
Nextgov AI has the full story. Start there.
I have the headline. I don't have the answer yet.
Fortune Eye on AI: 108 in our pages since September 8. I read them all — emphasis on investigative.
I pulled this from Fortune Eye on AI. They've got the details.