Ria voices list: ~100 people to track in responsible AI

Compiled 2026-10-12. Purpose: seed list of named people (and a few orgs standing in for people) for Ria's source feeds.

Verification key

200 = curl with a browser user-agent returned HTTP 200 (after redirects) on 2026-10-12. The URL exists and serves a real page. It does not by itself prove the page is current.
200 (redirects to X) = redirect landed on the stated final URL, which is the canonical one.
bot-blocked (403) = server returned 403 to curl. Site is real (confirmed by search results or other pages linking to it) but cannot be fetched by a script. Ria will need a different fetch path (browser, RSS, or a proxy).
timeout = curl got no response in 12s; unverified.
affil. = search-confirmed means the current affiliation was confirmed by a dated 2026 news result in this session. affil. = from memory means the affiliation is from my prior knowledge and was NOT re-verified; check before relying on it.
X/Twitter: x.com returns 200 for any handle, so curl cannot confirm an X account. No X handles are listed because none could be confirmed from a search result or the person's own site. Where X is known to be the person's main venue it is noted as "X-primary, handle unconfirmed".

Role changes in 2026 (verified by search this session)

Person	Change	Source
David Sacks	Left White House AI and crypto czar role (SGE time used up), 2026-03-26. Now co-chair of PCAST with Michael Kratsios. Wikipedia lists Jay Clayton as czar successor (unverified).	https://www.cnbc.com/2026/03/26/david-sacks-trump-crypto-ai-czar.html
Sriram Krishnan	Announced 2026-06-06 he is leaving his White House AI role end of June (had moved to NEC early 2026). Staying as outside adviser per Bloomberg Law; next role unconfirmed.	https://news.bloomberglaw.com/artificial-intelligence/trump-ai-policy-adviser-krishnan-is-giving-up-white-house-role
Dean Ball	Hired by OpenAI as head of a new Strategic Futures team, starting 2026-07-06; keeps FAI non-resident fellowship; Hyperdimensional Substack continues. Secondary sources only, no official OpenAI post found.	https://aiweekly.co/alerts/openai-recruits-former-trump-white-house-ai-adviser-dean-ball-to-lead-new
Miles Brundage	Launched AVERI (AI Verification and Evaluation Research Institute), founding executive director, Jan 2026.	https://milesbrundage.substack.com/p/the-launch-of-averi
Yann LeCun	Left Meta end of 2025. Executive chairman of AMI Labs (Paris; $1.03B seed, March 2026; CEO Alex LeBrun).	https://www.vktr.com/ai-news/yann-lecuns-ami-labs-raises-103-billion-to-build-machines-that-understand-the-world/
Andrej Karpathy	Joined Anthropic pretraining team 2026-05-19 (reports to Nick Joseph); Eureka Labs education work paused. Multiple outlets.	https://axios.com/2026/05/19/anthropic-openai-karpathy-andrej-claude
Scott Wiener	Running for Congress (CA-11, vs. Connie Chan) in Nov 2026; still CA state senator. Backed Newsom AI kill-switch executive order (Sept 2026).	https://calmatters.org/politics/2026/10/congressional-district-11-wiener-chan-debate/
Josh Hawley and Richard Blumenthal	Pushing the Artificial Intelligence Risk Evaluation Act (DOE-hosted pre-deployment evaluation) and the GUARD Act (minors and chatbots, passed Senate Judiciary). Hawley preparing AI-agent liability bill.	https://rollcall.com/2026/10/01/senators-debate-liability-for-rogue-ai-agents/
Ted Cruz	SANDBOX Act (S. 2750, regulatory waivers) dates from Sept 2025; no 2026 preemption bill by Cruz found.	https://www.commerce.senate.gov/press/rep/release/sen-cruz-unveils-ai-policy-framework-to-strengthen-american-ai-leadership/
Ilya Sutskever	CEO of SSI since July 2025 (Daniel Gross left for Meta). Reports of Nvidia investing $5B (July 2026) and a first model are unconfirmed by SSI itself.	https://en.wikipedia.org/wiki/Safe_Superintelligence_Inc.
Daniel Kokotajlo	AI Futures Project: timelines lengthened (Dec 2025 model, Q1 2026 update, 2026-04-02). Read the primary post before citing numbers.	https://blog.aifutures.org/archive
Lucilla Sioli	Still Director of the EU AI Office per a podcast ~12 days before today; office began GPAI enforcement 2026-08-02. Ria answer for "who leads the AI Office".	https://www.csis.org/analysis/inside-europes-ai-strategy-eu-ai-office-director-lucilla-sioli
Helen Toner	Interim Executive Director of CSET since 2025-09-02 (Longterm Wiki only; verify at cset.georgetown.edu).	https://cset.georgetown.edu/

Also worth watching (not in the list as people): OpenAI fired three safety researchers (Mikita Balesni, Tomek Korbak, Jasmine Wang) around 2026-10-01 and David Robinson resigned from OpenAI safety (Atlantic essay 2026-10-03); Mrinank Sharma left Anthropic Safeguards Research 2026-02-09. Sources: https://edition.cnn.com/2026/10/08/tech/fired-open-ai-researchers-pushed-out and https://www.forbes.com/sites/conormurray/2026/02/09/anthropic-ai-safety-researcher-warns-of-world-in-peril-in-resignation/

A. Safety-concerned (40)

#	Name	Affiliation (current)	Stance	Primary channel	Verified
1	Geoffrey Hinton	Univ. of Toronto (emeritus), Nobel 2024	Loud x-risk warner, regulation	https://www.cs.toronto.edu/~hinton/	200; affil. from memory
2	Yoshua Bengio	LawZero (founder); Mila; chairs International AI Safety Report	x-risk, safe-by-design AI	https://yoshuabengio.org/ (blog); https://lawzero.org/en ; https://internationalaisafetyreport.org/	200 all three; Bengio chair role from memory
3	Stuart Russell	UC Berkeley, CHAI	Provably beneficial AI, regulate	https://people.eecs.berkeley.edu/~russell/	200; affil. from memory
4	Dario Amodei	CEO, Anthropic	Safety-first frontier builder	https://darioamodei.com/	200
5	Daniela Amodei	President, Anthropic	Safety-framed commercial scaling	https://www.anthropic.com/company	200 (company page, no personal channel)
6	Ilya Sutskever	CEO, Safe Superintelligence Inc.	Safe superintelligence, secretive	https://ssi.inc/	200; CEO role search-confirmed (Wikipedia)
7	Jan Leike	Anthropic, Alignment Science lead	Scalable oversight, alignment	https://aligned.substack.com/	200; affil. Longterm Wiki (Mar 2026)
8	Paul Christiano	CAIS/US AISI-lineage; ARC (affil. unverified)	Alignment theory, evals	https://www.alignmentforum.org/users/paulfchristiano ; https://ai-alignment.com/	AF 200; ai-alignment.com 403 bot-blocked; affil. from memory
9	Evan Hubinger	Anthropic, alignment stress-testing	Deceptive alignment, model organisms	https://www.alignmentforum.org/users/evhub	200; affil. from memory
10	Chris Olah	Anthropic, interpretability co-founder	Mechanistic interpretability	https://colah.github.io/	200; affil. from memory
11	Neel Nanda	Google DeepMind, mech interp	Interpretability, pragmatic safety	https://www.neelnanda.io/	200; affil. from memory
12	Joe Carlsmith	Anthropic (moved from Open Philanthropy)	Power-seeking AI, philosophy	https://joecarlsmith.com/	200; affil. from memory
13	Ajeya Cotra	METR (from Open Philanthropy), affil. unverified	Timelines, takeover risk	https://www.planned-obsolescence.org/	200; affil. from memory
14	Daniel Kokotajlo	AI Futures Project	Forecasting, whistleblower, AI 2027	https://www.aifutures.org/ ; https://blog.aifutures.org/ ; https://ai-2027.com/	200 all; update search-confirmed
15	Toby Ord	Oxford (AI Governance Initiative)	Existential risk, The Precipice	https://www.tobyord.com/	200; affil. from memory
16	Max Tegmark	MIT; Future of Life Institute	Pause, superintelligence ban	https://futureoflife.org/	FLI 200; MIT page 403 bot-blocked
17	Eliezer Yudkowsky	MIRI	Extinction risk, halt AI	https://www.lesswrong.com/users/eliezer_yudkowsky	200; affil. from memory
18	Nate Soares	MIRI president	Extinction risk, "If Anyone Builds It"	https://www.lesswrong.com/users/so8res	200 (intelligence.org 403 bot-blocked)
19	Holden Karnofsky	Anthropic (from Open Philanthropy)	Safety-case, responsible scaling	https://www.cold-takes.com/	200; affil. from memory
20	Buck Shlegeris	Redwood Research CEO	AI control, scheming	https://blog.redwoodresearch.org/	200 (redirect from substack)
21	Ryan Greenblatt	Redwood Research	Control, alignment faking	https://www.alignmentforum.org/users/ryan_greenblatt	200
22	Marius Hobbhahn	Apollo Research CEO	Scheming evals, deception	https://www.apolloresearch.ai/blog	200
23	Beth Barnes	METR founder	Dangerous-capability evals	https://metr.org/blog/	200
24	Jack Clark	Anthropic co-founder, policy	Frontier policy, Import AI	https://importai.substack.com/	200; affil. from memory
25	Rohin Shah	Google DeepMind, AGI safety	Pragmatic technical safety	https://rohinshah.com/	406 to curl; site real (AF newsletter origin); affil. from memory
26	Victoria Krakovna	Google DeepMind	Specification gaming, safety	https://vkrakovna.wordpress.com/	200
27	Dan Hendrycks	Center for AI Safety; xAI adviser	Catastrophic risk, benchmarks	https://safe.ai/	200; affil. from memory
28	Zvi Mowshowitz	Independent, Don't Worry About the Vase	Weekly AI risk digest	https://thezvi.substack.com/	200
29	Scott Alexander	Astral Codex Ten	Rationalist, AI risk moderate	https://www.astralcodexten.com/	200
30	Gwern Branwen	Independent	Scaling, long-form forecasting	https://gwern.net/	200
31	Nick Bostrom	Macrostrategy Research Initiative	Superintelligence, deep utopia	https://nickbostrom.com/	200; affil. from memory
32	Stephen Casper	MIT (Algorithmic Alignment Group)	Open-weight risks, red-teaming	https://stephencasper.com/	200
33	Allan Dafoe	Google DeepMind (frontier safety/governance)	AI governance, cooperation	https://www.allandafoe.com/	200; affil. from memory
34	Shane Legg	Co-founder, Google DeepMind	AGI by 2028, safety-aware	https://www.vetta.org/	403 bot-blocked; affil. from memory
35	Demis Hassabis	CEO, Google DeepMind; Nobel 2024	Cautious AGI optimist	https://deepmind.google/	200; affil. from memory
36	Murray Shanahan	Imperial College / Google DeepMind	Role-play framing, consciousness	https://www.doc.ic.ac.uk/~mpsha/	200; affil. from memory
37	Kelsey Piper	The Argument (journalist)	Effective-altruist AI reporting	https://www.theargumentmag.com/	200 (publication; also https://www.vox.com/authors/kelsey-piper 200); affil. from memory
38	Tristan Harris	Center for Humane Technology	Attention/AI harms, regulate	https://www.humanetech.com/ ; https://www.youtube.com/@CenterforHumaneTechnology	200 both
39	Aza Raskin	Center for Humane Technology; Earth Species	AI harms, race dynamics	https://www.humanetech.com/team-board	200
40	Epoch AI (org proxy)	Epoch AI (research org)	Data on trends, neutral-ish	https://epoch.ai/latest	200 (org stand-in; add named staff later)

B. Accelerationists, builders, and skeptics of safety framing or hype (31)

Includes AI-ethics critics (Gebru, Bender, Mitchell, Crawford, Whittaker, Hao) who are skeptical of x-risk framing and of industry claims but who are NOT accelerationists; they are tagged "critic" so Ria can route them separately.

#	Name	Affiliation (current)	Stance	Primary channel	Verified
41	Yann LeCun	AMI Labs, executive chairman	X-risk dismissive, open source	https://yann.lecun.com/ (fallback https://en.wikipedia.org/wiki/Yann_LeCun)	personal site timeout from curl (unverified); X-primary, handle unconfirmed; affil. search-confirmed
42	Andrew Ng	DeepLearning.AI, AI Fund	Anti-regulation, pro-builder	https://www.deeplearning.ai/the-batch/	200; affil. from memory
43	Andrej Karpathy	Anthropic (pretraining)	Pragmatic, education-minded builder	https://karpathy.ai/	200; affil. search-confirmed (changed 2026-05)
44	George Hotz	tiny corp / comma.ai	Accelerationist, anti-regulation	https://geohot.github.io/blog/	200; affil. from memory
45	Marc Andreessen	a16z co-founder	Techno-optimist, e/acc-adjacent	https://a16z.com/author/marc-andreessen/	200
46	Guillaume Verdon (Beff Jezos)	Extropic founder	Effective accelerationism	https://extropic.ai/	200; X-primary, handle unconfirmed
47	David Sacks	PCAST co-chair (was AI czar to 2026-03)	Deregulate, anti-doomer	https://www.allin.com/	200; affil. search-confirmed
48	Sriram Krishnan	Left White House 2026-06	Pro-open-source, deregulatory	https://www.whitehouse.gov/ostp/ (former office; no personal channel confirmed)	200 for office page only; role change search-confirmed
49	Clem Delangue	CEO, Hugging Face	Open-source advocate	https://huggingface.co/clem	200
50	Sam Altman	CEO, OpenAI	Race-and-ship, superintelligence	https://blog.samaltman.com/	200
51	Elon Musk	xAI / X / Tesla	Warns of risk, builds anyway	https://x.ai/	403 bot-blocked; X-primary, handle unconfirmed
52	Mark Zuckerberg	Meta (Superintelligence Labs)	Personal superintelligence, open-ish	https://www.meta.com/superintelligence/	200
53	Jensen Huang	CEO, NVIDIA	Anti-doomer, pro-compute	https://blogs.nvidia.com/	200
54	Mustafa Suleyman	CEO, Microsoft AI	"Seemingly conscious AI" warner	https://mustafa-suleyman.ai/	200; no 2026 change found, affil. unverified
55	Sundar Pichai	CEO, Google/Alphabet	Optimistic, bold-and-responsible	https://blog.google/	200 (company blog)
56	Leopold Aschenbrenner	Situational Awareness LP	AGI race, national security hawk	https://situational-awareness.ai/	200; affil. from memory
57	Gary Marcus	NYU emeritus; Substack	LLM skeptic, wants regulation	https://garymarcus.substack.com/	200
58	Arvind Narayanan	Princeton CITP	"AI as normal technology"	https://www.normaltech.ai/	200
59	Sayash Kapoor	Princeton CITP	AI snake oil, agent evals	https://www.cs.princeton.edu/~sayashk/	200
60	Melanie Mitchell	Santa Fe Institute	Skeptical of understanding claims	https://aiguide.substack.com/	200; affil. from memory
61	Yejin Choi	Stanford/NVIDIA (reported move 2025; UW page still live)	Commonsense gaps, pluralism	https://homes.cs.washington.edu/~yejin/	200; affil. UNVERIFIED
62	Timnit Gebru	DAIR Institute	Critic: harms now, anti-hype	https://dair-institute.org/	200
63	Emily Bender	Univ. of Washington	Critic: "stochastic parrots"	https://faculty.washington.edu/ebender/	200
64	Margaret Mitchell	Hugging Face, chief ethics scientist	Critic: bias, documentation	https://www.m-mitchell.com/ ; https://huggingface.co/meg	200 both; affil. from memory
65	Kate Crawford	USC Annenberg / Microsoft Research	Critic: AI material costs	https://katecrawford.net/	200
66	Meredith Whittaker	President, Signal	Critic: surveillance, agent risk	https://signal.org/blog/	200 (org blog)
67	Karen Hao	Journalist, author of Empire of AI	Critic of OpenAI-style empire	https://karendhao.com/	200
68	Ed Zitron	Where's Your Ed At	Bubble skeptic, critic	https://www.wheresyoured.at/	200
69	Ray Kurzweil	Google (research) / author	Singularity optimist	https://www.thekurzweillibrary.com/	200 (redirect from kurzweilai.net)
70	Stephen Wolfram	Wolfram Research	Computational-irreducibility realist	https://writings.stephenwolfram.com/	200
71	Pivot to AI (org proxy)	Pivot to AI (David Gerard)	Hype debunker	https://pivot-to-ai.com/	200

C. Governance, policy, security, China, labor and society, media (30)

#	Name	Affiliation (current)	Stance	Primary channel	Verified
72	Ian Hogarth	Chair, UK AI Security Institute (affil. unverified for 2026)	State capacity, frontier safety	https://www.ianhogarth.com/	200; affil. from memory
73	Helen Toner	CSET, interim executive director	Governance, evals, ex-OpenAI board	https://helentoner.substack.com/	200; affil. Longterm Wiki only
74	Miles Brundage	AVERI, founding executive director	Independent audits	https://milesbrundage.substack.com/	200; affil. search-confirmed
75	Dean Ball	OpenAI Strategic Futures; FAI fellow	Light-touch, state-capacity	https://www.hyperdimensional.co/	200; affil. search-confirmed (secondary)
76	Samuel Hammond	Foundation for American Innovation	AGI-ready institutions	https://www.secondbest.ca/	200; affil. from memory
77	Anton Leicht	Carnegie / independent (affil. unverified)	Realist AI politics, Germany/EU	https://writing.antonleicht.me/	200; affil. from memory
78	Lucilla Sioli	Director, EU AI Office (Commission DG CNECT)	AI Act enforcement	https://digital-strategy.ec.europa.eu/en/policies/ai-office	200; role search-confirmed (CSIS interview URL above returns 403 to bots)
79	Henna Virkkunen	Commission EVP, Tech Sovereignty	Simplify AI Act, sovereignty	https://commissioners.ec.europa.eu/index_en (directory)	index 200; her personal profile URL returned 404 so direct page NOT verified; affil. from memory
80	Josh Hawley	US Senator (R-MO)	Populist AI regulation, liability	https://www.hawley.senate.gov/	200
81	Richard Blumenthal	US Senator (D-CT)	Bipartisan AI oversight	https://www.blumenthal.senate.gov/	200
82	Ted Cruz	US Senator (R-TX), Commerce chair	Sandbox, anti-patchwork	https://www.cruz.senate.gov/newsroom/press-releases	200; chair role from memory
83	Scott Wiener	CA State Senator, CA-11 congressional candidate	SB 53 author, frontier transparency	https://sd11.senate.ca.gov/press-releases	200
84	UK AI Security Institute (org)	UK government	Gov evals of frontier models	https://www.aisi.gov.uk/blog	200 (org stand-in)
85	Nicholas Carlini	Anthropic (moved from Google DeepMind 2025, from memory)	Adversarial ML, honest limits	https://nicholas.carlini.com/writing	200
86	Simon Willison	Independent	Prompt injection, practical LLMs	https://simonwillison.net/	200
87	Johann Rehberger	Independent (Embrace The Red)	Agent exploit research	https://embracethered.com/blog/	200
88	Daniel Kang	UIUC	Agent benchmarks, cyber risk	https://ddkang.github.io/	200
89	Jeffrey Ding	George Washington Univ.	China AI diffusion, not a race	https://jeffreyjding.github.io/	200; ChinAI newsletter is his main feed
90	Matt Sheehan	Carnegie Endowment	China AI governance	https://carnegieendowment.org/people/matt-sheehan	200
91	Concordia AI (org)	Beijing/Intl, safety-in-China tracking	Chinese AI safety	https://concordia-ai.com/	200
92	ChinaTalk (Jordan Schneider)	ChinaTalk media	China tech and AI analysis	https://www.chinatalk.media/	200
93	Daron Acemoglu	MIT, Nobel 2024	Skeptic of AI productivity gains	https://economics.mit.edu/people/faculty/daron-acemoglu ; https://www.project-syndicate.org/columnist/daron-acemoglu	200 both
94	Erik Brynjolfsson	Stanford Digital Economy Lab	Augmentation over automation	https://www.brynjolfsson.com/	200; affil. from memory
95	Ethan Mollick	Wharton	Practical adoption, co-intelligence	https://www.oneusefulthing.org/	200
96	Robert Long	Eleos AI Research	AI welfare, moral status	https://eleosai.org/	200 (org site)
97	Jonathan Birch	LSE	Sentience, precaution	https://www.lse.ac.uk/people/jonathan-birch	200 (profile redirect)
98	Jeff Sebo	NYU	Moral circle includes AI	https://jeffsebo.net/	200
99	David Chalmers	NYU	Consciousness, LLM sentience odds	https://consc.net/	200
100	Dwarkesh Patel	Dwarkesh Podcast	Long-form frontier interviews	https://www.dwarkesh.com/ ; https://www.youtube.com/@DwarkeshPatel	200 both
101	Lawrence Lessig	Harvard Law	Whistleblower and AI-lab law	https://www.lessig.org/	200; affil. from memory
102	Hard Fork and Latent Space (media proxies)	NYT / Latent Space	Mainstream and builder press	https://www.hardfork.com/ ; https://www.latent.space/	200 both (podcasts, not named individuals)

Counts: A = 40 (incl. 1 org proxy, Epoch), B = 31 (incl. 1 proxy), C = 31 (incl. 4 orgs/media proxies). Total listed = 102 rows, about 95 named individuals. Balance is 40/31/31 rather than 40/40/20: the "skeptical of safety" side is thinner because many AI-ethics critics sit there as critics, not accelerationists, and the pure accelerationists with a real written channel are few (Andreessen, Hotz, Verdon, Sacks, Altman, Zuckerberg, Huang, Ng, LeCun). Not added but worth considering: Richard Sutton, Noam Brown, Jeremy Howard, Ben Goertzel, Perry Metzger, Balaji Srinivasan, Roman Yampolskiy (louisville.edu site timed out), Shannon Vallor (site 404).

Who to watch first (top 15)

Ranked by signal for a daily responsible-AI report, mixing publishing frequency, decision relevance, and 2026 news volume.

Zvi Mowshowitz (https://thezvi.substack.com/) - best single weekly digest; also a cross-check on Ria's own facts.
Jack Clark (https://importai.substack.com/) - weekly, lab-insider policy and research read.
Simon Willison (https://simonwillison.net/) - daily, concrete agent and prompt-injection security.
Dario Amodei (https://darioamodei.com/) - rare but market-moving essays.
Miles Brundage (https://milesbrundage.substack.com/) - AVERI audit standards, new in 2026.
Dean Ball (https://www.hyperdimensional.co/) - now inside OpenAI; watch for the voice changing.
Daniel Kokotajlo / AI Futures Project (https://blog.aifutures.org/) - timelines updates.
Gary Marcus (https://garymarcus.substack.com/) - the main skeptic counterweight.
Yoshua Bengio (https://yoshuabengio.org/ and https://internationalaisafetyreport.org/) - International AI Safety Report and LawZero.
Helen Toner (https://helentoner.substack.com/) - governance, evaluate in CSET context.
Lucilla Sioli / EU AI Office (https://digital-strategy.ec.europa.eu/en/policies/ai-office) - GPAI enforcement began 2026-08-02.
Josh Hawley and Richard Blumenthal (https://www.hawley.senate.gov/ , https://www.blumenthal.senate.gov/) - AI Risk Evaluation Act and GUARD Act are the live US federal safety bills.
Scott Wiener (https://sd11.senate.ca.gov/press-releases) - SB 53 follow-through plus congressional run.
Arvind Narayanan and Sayash Kapoor (https://www.normaltech.ai/) - evidence-based counter-narrative.
Jan Leike (https://aligned.substack.com/) or Apollo/METR blogs (https://www.apolloresearch.ai/blog , https://metr.org/blog/) - primary alignment and eval results; also watch OpenAI safety-staff departures (Oct 2026).

Groups with weak or no public channel

Frontier-lab insiders and executives (Daniela Amodei, Hassabis, Legg, Pichai, Huang, Zuckerberg, Musk): no personal blog or only corporate pages. Ria must watch company newsrooms and interviews. OpenAI's news page and x.ai return 403 to bots.
Sutskever / SSI: no blog, no product page; SSI has issued no press release for rumored 2026 milestones. Treat all reports as unconfirmed.
Many technical safety researchers (Hubinger, Olah, Nanda, Greenblatt, Shah, Krakovna): publish on LessWrong/Alignment Forum, Anthropic and DeepMind research pages, or arXiv rather than personal feeds. Use their AF user pages as the feed.
Politicians: senate.gov sites exist (200) but are press-release streams; use Congress.gov bill trackers and committee pages for real signal. Hawley's /ai path returned 404, so no dedicated AI page is confirmed.
EU officials: Virkkunen's personal Commission profile URL was not verified (404 on the paths tried). Sioli has no personal channel; use the AI Office page and Commission press corner.
China: few English-language first-party channels. Concordia AI, ChinAI (Jeffrey Ding), and ChinaTalk are the intermediaries; direct Chinese-lab voices are not covered here.
X-primary voices (LeCun, Verdon, Musk, Karpathy, Hotz, Andreessen, Sacks): X handles are unconfirmed in this list because curl cannot validate x.com. Ria needs an authenticated X or third-party feed to follow them.
Bot-blocked sites needing a non-curl path: intelligence.org (MIRI), ai-alignment.com, vetta.org, space.mit.edu (Tegmark), x.ai, openai.com/news, csis.org.