Physicists are warning against Effective Altruists spreading a “mind virus” over Silicon Valley, promoting a “doomer” view of Artificial Intelligence.
asserted
Altruists → warn → Intelligence
Calls to regulate AI have been growing from industry leaders Anthropic and OpenAI, who warn the technology they are building could kill us all.
uncertain
they → regulate → us
However, those who list themselves as AI accelerationists feel such companies are bluffing in a bid to safeguard the advantages they have gained after spending billions of dollars creating their AI models.
asserted
they → list → models
“I don’t think AI is going to kill us all.
asserted
AI → think → us
I think that sort of thinking is more dangerous than the AI itself,” Guillaume Verdon, the physicist who founded the Effective Accelerationism movement, told The Post.
asserted
who → think → Post
“If you ask any of these doomers how it’s supposed to kill us all, they can never flesh out all the details …
asserted
they → ask → details
I’m a mathematician.
asserted
I → ’m → ?
Verdon pointed out how the practice of Effective Altruism — a philosophy which believes a certain circle of enlightened people committed to doing the most good are the only ones to be trusted — has “culturally, ideologically compromised all these big labs.”
asserted
circle → point → labs
“Someone has to stand up to this complex,” he said, noting he has endured pushback he blames on AI “doomers.”
asserted
he → have → doomers
“They think I’m Voldemort,” he added, referring to the Harry Potter villain.
asserted
he → think → villain
“They are obsessed with ‘Harry Potter.’
asserted
They → obsess → Potter
Anthropic CEO Dario Amodei published “We Must Pace the Frontier,” earlier this month urging companies to slow advances in their most powerful systems, embed outside evaluators inside their labs and work with the government on regulation.
asserted
We → publish → regulation
But Verdon noted the proposed regulators are all friends with the CEOs and higher ups at the biggest AI companies.
asserted
regulators → note → companies
“They propose the third-party auditors that are all from the same ideological school of thought.
asserted
that → propose → thought
,” he said.
asserted
he → say → ?
“They’re just gonna be a sort of cartel with fake checks and balances,” he charged.
asserted
he → go → checks
Anthropic did not respond to a request for comment from The Post.
asserted
Anthropic → respond → Post
Government regulation as well as partisan regulators aligned with Anthropic and OpenAI would be huge obstacles for smaller companies — or those who don’t adhere to the same philosophy.
asserted
who → align → philosophy
“The AI CEOs can only do so much,” Verdon said, describing Anthropic CEO Amodei as “a very thoughtful guy.”
asserted
Verdon → do → guy
“I don’t think they’re necessarily bad people,” he said of people he knows at the company.
asserted
he → think → company
“I think they are just so deep in this ideology they can’t see clearly.”
asserted
they → think → ideology
President Trump and senior government officials have rejected calls for new guardrails for AI companies, telling them there is nothing stopping them from regulating themselves.
The public remains uneasy.
asserted
public → reject → themselves
A Pew Research Center survey of 3,488 adults in June found 52% were more concerned than excited about AI’s growing role in daily life.
asserted
% → find → life
Verdon says his Effective Accelerationism movement is completely decentralized.
asserted
movement → say → ?
Its zeal for rapid technological progress has found a champion in venture capitalist Marc Andreessen’s “Techno-Optimist Manifesto.”
asserted
zeal → find → Manifesto
“We don’t have billions of dollars of shadow funding trying to influence things,” Verdon said, tongue-in-cheek.
asserted
Verdon → have → cheek
He advocates for more competition, more open-source AI models and freedom to build.
asserted
He → advocate → competition
His startup, Extropic, is developing energy-efficient computing chips.
asserted
startup → develop → chips
In July the company announced a nonbinding letter of intent for up to $75 million in Commerce Department funding.
asserted
company → announce → funding
Not every builder is sold.
asserted
builder → sell → ?
Bojan Tunguz, a data scientist and founder of Tampa-area startup TabulAI, calls himself “adjacent” to e/acc.
asserted
Tunguz → call → acc
“I don’t believe that acceleration at any cost is an absolute good,” he told The Post.
asserted
he → believe → Post
“I do think that some cap might be necessary at some point and some pacing might be justifiable.”
uncertain
pacing → think → point
Noting the rapid pace of growth, he said today’s AI tools already let him build things that were not possible “even six months ago.”
asserted
that → note → things
He welcomes open models as a reliable fallback if big labs restrict access to their tools.
asserted
labs → welcome → tools
But he rejects the idea cornering the market is their motive.
asserted
cornering → reject → market
Asked about Verdon’s “mind virus” claims, Tunguz took a broader view.
uncertain
Tunguz → ask → view
“I think we all have our own mind viruses — our own perspectives, our own favorite worldviews,” he said.
asserted
he → think → viruses
“I think we are all to some extent accelerationists.
asserted
we → think → extent
The Aporia
