I made a three-minute documentary about what (usually) happens when a company brings in AI. Most of it is true, which is the problem.

I built a magazine staffed by AI agents: six columnists, a full production pipeline, and an AI Editor-in-Chief. I kept three jobs. Reading before it goes live, publishing, and my own column. Here's how the whole operation runs: varops.com/how-we-run-varops…

The most is talent + drive.

No thank you

Reasoning effort growing up

Hi @bot team - time to create an Apple Watch app! @poteto @mattyp @elonmusk @cursor_ai @SpaceXAI

Lots of fine prints but still… We're getting close to being able to “just throw compute” at most (well defined) problems. What a time to be alive.

97% of parallel agent runs were interfered with by exploits. Here's how to measure parallel-agent interference on a repo before buying a tool to prevent it ↓ varops.com/how-to-measure-pa…

~6 months... I'm betting we'll see "close enough" indie and Chinese clones before that.

Opus 5.5, GPT-6 Sol, Grok 4.7, and Jev all came out in a span of one week. They are all strong releases – but none of them is where the edge comes from for *understanding* a company. Let's talk about it nitter.net/i/broadcasts/1aJbdEpjg…

LIVE: Old School / New Tech nitter.net/i/broadcasts/1aJbdEpjg…

Every change lands on a live staging environment before production. We look at it and give one of three answers: ship it, fix this first, or don't ship it. Within one business day. In writing, on a record. automaze.io/services/cto-as-…

We’ll start pacing AI tomorrow.

The original @bot

Time to play! @claudeai Opus 5.5 is out! Promises to perform at the level of Claude Fable 5.1 for most tasks and costs 40% less to run than Opus 5.

I tested both Jev and Laya against @muxi_ai's pre-request gates and we're not switching. Laya is the open-weights version of Jev, @typesafeai's System One model, a classifier that returns typed decisions instead of text. The pitch is speed and no hallucinations, because nothing gets generated. MUXI already runs 11 of these decisions before any request reaches an agent. Is this actionable, does it need clarification, is it a recall question, is the user trying to stop. They run on a 95 MB ONNX embedding model with hand-picked examples per gate. 100% on all 11 eval sets, about 50 ms for all. I pointed Laya at the two decisions we still hand to gpt-4o-mini. Same fixtures, same labels, incumbent vs Jev vs Laya. Credential detection: 96.7% / 52.2% / 54.4% Complexity within one point: 80.0% / 27.5% / 10.0% Routing accuracy: 71.4% / 81.1% / 33.3%. Prompt-injection recall: 10/10 / 3/10 / 10/10 False positives on safe messages: 2.3% / 14.0% / 93.0% p50 latency: 866 ms / 127 ms / 69 ms Jev/Laya is seven to twelve times faster. they also misses 7 of 10 attacks, or flags 93% of normal requests as attacks, depending on which checkpoint you pick. Routing is the security check in MUXI. Fast and wrong on that gate is not a trade. The reframe I took away is that Jev's pitch is right and its model is beside the point. Decide before you generate, with typed answers, is the correct design. We already had it. A small embedding model with a dozen examples per gate beat a 300 to 400M parameter decision model on our data, because our gates are generic intent questions and that is exactly the regime embeddings are good at. What we're doing instead is keeping the Jev-shaped interface, typed questions in and typed answers out, and running it over our own model. All 11 gates now share one embed per request. And the complexity call to the LLM only fires when a request actually needs decomposing. muxi.org

One component in our AI system, codenamed 'Varys', analyzes company processes in real time, identifies bottlenecks, and automates tasks to improve efficiency. It was so fun to finally talk about it :)

Jev

Models are replaceable compute. Your company’s understanding shouldn’t be.