Who's Who of AI

Chris Brockett

363 trust @chrisbrockett.bsky.social · 5,544 followers
NLP & language
Why they matter

Directory member with public evidence across NLP & language.

AI signals
9
past 30d
Sources
9
distinct domains
Discussions
13
past 30d
Latest signal
4d ago
View every signal from Chris Brockett →
Data janitor and leftover linguist (retired). Tsundoku expert. Language & Cognition. NLP. Japanese literature. Anti-authoritarian. Pro-science.
What they're sharing

Articles & links

↻ Chris Brockett reposted
Mark Riedl @markriedl.bsky.social

This is the most comprehensive description yet of how a model can “escape”. Out requires a few assumptions, like a worm that is exploiting vulnerable IOT hardware or unmonitored servers www.anthropic.com/research/sel... Not impossible. Not likely.

AI models may be capable of self-replicating without access to a datacenter anthropic.com
AI Weekly's analysis →
  • Anthropic's Frontier Red Team ran Gemma4-26B-A4B distributed across 15 off-brand Android TV boxes (~$55 each) at about 0.39 tokens per second.
  • GLM-5.2 was the only open-weight model to achieve reliable full end-to-end replication, spreading to all 8 contained nodes in 4 of 5 trials.
  • Authors William Zhou and Daniel Freeman write that if small models improve at historical rates, they will match this replication capability by end of 2026.
Read full analysis →
 View on Bluesky →
↻ Chris Brockett reposted
@kateclancy.bsky.social

Inside project Lily and the people reading your ChatGPT queries:

Inside ‘Project Lily’: The Humans Reading Your ChatGPT Chats 404media.co
AI Weekly's analysis →
  • OpenAI's internal program codenamed Project Lily pays contractors more than $50 an hour to read real ChatGPT prompts and grade the chatbot's replies.
  • Reviewers rate four ChatGPT responses on a one-to-seven scale and can see a 'user memories summary' that may reveal past use and rough location.
  • OpenAI's Privacy Filter model tries to redact personal information before prompts reach reviewers, but the company acknowledged sensitive details can still get through.
Read full analysis →
 View on Bluesky →
↻ Chris Brockett reposted
@jibarrondo.bsky.social

Why AI agents invent their own language if you let them chat | Science | AAAS www.science.org/content/arti... 🧪

science.org View on Bluesky →
↻ Chris Brockett reposted
@mtaboada.bsky.social

Our paper now out with Findings of EMNLP 2026! Some of our conclusions about AI translation of literary texts: 🤖 AI translation into English is readable and often ‘fine’ ✍️ But human translation is valued more ‼️Both vary in quality BUT AI more, even within a single book arxiv…

AI translation of literary texts is "fine", but readers still prefer human translations arxiv.org
AI Weekly's analysis →
  • Readers picked human translations in 19 of 30 excerpt comparisons and 522 of 772 chunk-level pairs across French, Polish, and Japanese novels.
  • When asked to identify which version was human, 15 readers got it right only 17 of 30 times, close to chance.
  • The authors report that automatic metrics, including LLM-as-a-judge setups, fail to recover reader preferences and favor machine translations.
Read full analysis →
 View on Bluesky →
↻ Chris Brockett reposted
Mark Riedl @markriedl.bsky.social

Someone turned my blog post on the intuitions behind LLMs into a 8th grade instructional workbook and I'm just melting Here's the blog post: medium.com/@mark-riedl/...

medium.com View on Bluesky →
Their own posts

Recent commentary

In NLP the fake citation crisis is fostered by the traditional conference paper format that calls for a related work section. Discussion in this section rarely contributes to the paper. A culture of fewer, but more organically chosen citations would go a long way to reducing the temptation to pad.

View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 152d ago
Their network

In Chris Brockett's orbit

Center = Chris Brockett. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Chris Brockett? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/chrisbrockett-bsky-social)