RC RANDOM CHAOS

AI

13 posts

OpenCode Console breach exposed stored provider API keys
Article

OpenCode Console breach exposed stored provider API keys

An intruder reached a legacy OpenCode system on October 16, possibly exposing provider keys for 1.4% of Console users. What to revoke and check now.

Beam's weights are still pending
Article

Beam's weights are still pending

Reflection's Beam is a 501B-parameter sparse open-weight model, but its weights and technical artifacts are still pending release.

The benchmark score is the number to trust least
Article

The benchmark score is the number to trust least

Claude Opus 5.5 scores 58 on the Intelligence Index, but the cost, verbosity and latency numbers behind it matter more.

One write past the chunk, read-write on the repo
Article

One write past the chunk, read-write on the repo

How Hacktron chained a libheif heap overflow and an OpenAI SSO misconfiguration to reach OpenAI's internal repo in under 72 hours, with AI doing the exploit work.

What distillation leaves behind
Article

What distillation leaves behind

Y Combinator's Garry Tan opposes regulating AI distillation and wants US open-weight labs to distill American frontier models through the front door.

Distillation makes fast followers, never frontier leaders
Article

Distillation makes fast followers, never frontier leaders

Garry Tan wants US open-weight labs to distill frontier models legally. Distillation copies a teacher, so it follows the frontier but can't lead it.

The model kept no receipts
Article

The model kept no receipts

A mathematician presents evidence that OpenAI may have trained its Astra model on his private work on a conjecture it was then credited with solving.

SWE-2 is the wrong model for almost everything
Article

SWE-2 is the wrong model for almost everything

Cognition's SWE-2 ships in medium, high, and max tiers; by its own numbers medium is the right pick for almost every coding task.

Same task, different bets
Article

Same task, different bets

A study of nearly 17,000 coding-agent sessions found agents converge on the same tool picks across personas and codebases, with full traces published.

AI could make you a stranger to the machines you depend on
Article

AI could make you a stranger to the machines you depend on

An essay on losing the familiarity of the struggle that taught engineers how their machines work, as AI absorbs the fiddly work that used to teach it.

The demo passed. Two weeks later, the queue filled.
Article

The demo passed. Two weeks later, the queue filled.

Charity Majors argues cheap AI code turns code into a cache and shifts engineering work to validation, observability and production discipline.

Stanford teaches LLMs by making you build one
Article

Stanford teaches LLMs by making you build one

Stanford CS336 teaches language models by having students build the training stack end to end, from tokenizers to data and alignment.