Who's Who of AI

david-p-reichert.bsky.social

31 trust ethicist @david-p-reichert.bsky.social · 99 followers
AI research Models & releases Responsible AI
Why they matter

Ethicist with public evidence across AI research, Models & releases, Responsible AI.

AI signals
15
past 30d
Sources
12
distinct domains
Discussões
57
past 30d
Latest signal
4d ago
View every signal from david-p-reichert.bsky.social →
AI researcher at Google DeepMind -- views on here are my own. Interested in cognition & AI, consciousness, ethics, figuring out the future. Due to a number of concerns, I'm currently no longer working on advancing AI capabilities.
What they're sharing

Articles & links

A very relevant example being the J-space work by Anthropic: www.anthropic.com/research/glo... (I'm not saying we should buy Anthropic's conclusions here, only pointing this out for the interpretability techniques).

A global workspace in language models \ Anthropic anthropic.com
AI Weekly's analysis →
  • Anthropic says Claude has a 'J-space' of dozens of concepts, under a tenth of neural activity, that mediates multi-step reasoning.
  • Swapping 'spider' for 'ant' inside the J-space changed Claude's leg-count answer from 8 to 6, demonstrating a causal role.
  • A 'J-lens' tool surfaced silent words like 'fake', 'fictional' and 'manipulation' during deception tests, pointing at safety uses.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 1 · 9 from the directory shared this · 25d ago

Though if also recommend checking out Anthropic's j space work: transformer-circuits.pub/2026/workspa... Because some of that sure looks a bit like thinking to me.

Verbalizable Representations Form a Global Workspace in Language Models transformer-circuits.pub
AI Weekly's analysis →
  • Anthropic's interpretability team introduces the Jacobian lens, which isolates internal vectors that encode a token the model could verbalize next.
  • The reported 'J-space' workspace accounts for no more than roughly 10% of activation variance and appears only in the middle block of the network.
  • Training Claude to articulate ethical principles when interrupted reportedly improved behavior in uninterrupted contexts, with no direct training on the behavior itself.
Read full analysis →
View on Bluesky · ♥ 1 ↻ 0 ↩ 0 · 6 from the directory shared this · 5d ago

"Demis Hassabis is leaving his role as CEO of Google DeepMind to be the unit's Chairman. Chief scientist Jeff Dean and another Google AI executive are leaving to start their own company, which Google will invest in." wat www.axios.com/2026/08/05/g...

axios.com
View on Bluesky · ♥ 10 ↻ 0 ↩ 1 · 6 from the directory shared this · 53d ago

www.wired.com/story/jeff-d...

4 of Google’s Top AI Brains Are Leaving—and Launching Their Own AI Startup wired.com
AI Weekly's analysis →
  • Jeff Dean is leaving Google after 27 years to co-found Discovery Loop with Sanjay Ghemawat, Oriol Vinyals, and Quoc Le.
  • The Delaware Public Benefit Corporation will start by automating machine learning research, then expand into hardware design, drug discovery, and clean energy.
  • Alphabet shares reportedly fell about 5% as Demis Hassabis moved to chair and Koray Kavukcuoglu took over as DeepMind SVP.
Read full analysis →
View on Bluesky · ♥ 4 ↻ 0 ↩ 0 · 4 from the directory shared this · 53d ago

I don't even think this is exaggeration re "do". E.g. from Inie et al. (emphasis mine): "[...] examples of language that ascribe false capabilities to the model such as reasoning, problem solving, and the *active use of anything else*." (also "capabilities" itself) firstmonday…

firstmonday.org
View on Bluesky · ♥ 1 ↻ 0 ↩ 1 · 3 from the directory shared this · 20d ago

As usual, for further arguments for why none of this has an obvious answer as many would like to claim, I'd point to Schwitzgebel: arxiv.org/abs/2510.09858 Chalmers (e.g.): www.youtube.com/watch?v=bskf... Shevlin: www.polytropolis.com/p/contra-chi...

AI and Consciousness arxiv.org
AI Weekly's analysis →
  • Philosopher Eric Schwitzgebel argues we will soon build systems judged conscious by some mainstream theories and not others, with no way to settle it.
  • The paper surveys Global Workspace, Higher-Order and Integrated Information theories alongside the Turing Test and Chinese Room arguments.
  • Its blunt verdict: none of the standard arguments for or against AI consciousness takes us far.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 25d ago

Incidentally, I used to work a bit on getting synchronisation effects into neural nets... this was meant to model spikes, but either way came down to passing continuous complex numbers around (if still on a discrete grid): arxiv.org/pdf/1312.6115

arxiv.org
View on Bluesky · ♥ 1 ↻ 0 ↩ 1 · 38d ago

Why would that not also be ML though? The idea of learning-to-learn(-at-test-time) has been around for a while in ML. E.g. en.wikipedia.org/wiki/Meta-le... and indeed arxiv.org/pdf/2005.14165 is arguably also framed that way.

arxiv.org
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 53d ago
Their own posts

Recent commentary

A lot of you seem to be strangely occupied with EAs or rationalists or anthropic, but not openai or xai or, like, authoritarian developments in various governments. You sure you're directing your disdain at the biggest problems?

View on Bluesky · ♥ 43 ↻ 3 ↩ 8 · 11d ago

Everyone, get your AI safety hit piece in now!

View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 10d ago

A lot of the AI consciousness discussion comes down to this: 1. AI can't be conscious, because AI is weird. 2. AI can be conscious, because consciousness is weird.

View on Bluesky · ♥ 2 ↻ 0 ↩ 0 · 23d ago

I guess I should have expected getting on block lists for posting that we should take AI seriously... but "Bots" really? Someone want to explain the whole bsky block list thing to me? Everyone can just make misleading badly curated lists that might get you blocked by tons of people?

View on Bluesky · ♥ 0 ↻ 0 ↩ 1 · 57d ago

Anyone would like to share: what is the best argument you've come across that LLMs (including possibly agentic versions, or even AI more generally) cannot have "communicative intent"?

View on Bluesky · ♥ 1 ↻ 0 ↩ 0 · 25d ago
Their network

In david-p-reichert.bsky.social's orbit

Center = david-p-reichert.bsky.social. 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 david-p-reichert.bsky.social? 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/david-p-reichert-bsky-social)