Slowing AI development
“Frontier AI development should slow down relative to its current default trajectory”
28 people · 15 agree · 6 disagree · 6 no clear side · 3 unclear · 42 views · 2026-06-05 to 2026-10-23
Who stands where on the statement above, from their own words. A person's side is shown only where two separate readings of the quote agree; otherwise they are under "Side unclear". This page does not say who is right.
Agree 15 people
- ✓Center for Humane Technology
the companies have been badly incentivized to build a God in the box that then they then self-admit that they don’t know how to control.
if we paused with the current power of AI that we have right now, you would still get 10 years of advancement, economic growth, productivity, GDP growth
it’s time for a verified red line on using compute for recursive self-improvement.
We need to break this spell of inevitability that follows each fresh cycle of model releases and uncontrollable harms.
Standing in a similar moment - with urgent timelines and mounting evidence of AI’s risks
- ✓Dario Amodei
AI developers should make a conscious effort to slow down, rather than continue racing to push out newer and more powerful frontier models.
Some forms of coordination that would be impactful for pacing are legally challenging, and will require government support,
- ✓Jakub Pachocki
He proposed that companies should voluntarily pause their work, or at least slow down, until the industry can establish concrete safety standards around the technology.
no lab has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer.
- ✓President Halla Tómasdóttir
I am a huge believer in pausing
the power and speed of technology surpasses our ability to find wisdom and come together around how to govern this.
- ✓Import AI
We will need to move extremely carefully here, because otherwise we could lose control of the technology.
we are going to need to work together to create unusual and serious rules for this technology and the companies building it, and to exercise the hardest and I think most noble quality of people – restraint.
- ✓Zvi Mowshowitz
We urgently need to slow down, relative to what will happen by default.
I think we should be trying to slow it down.
- ✓Bernie Sanders
an immediate pause on advanced AI development, and a permanent ban on superintelligence
- ✓Anthropic spokesperson
the world would benefit from the industry adopting a lawful, verifiable way to work together to pace how we release powerful models.
- ✓Sam Altman
I agree with Dario that we need to pace the frontier.
- ✓Elon Musk
Dario is right
- ✓Thomas Larsen
It’s going to happen. It’s going to happen unless we take deliberate steps to stop it.
- ✓roon
there have not been fundamental advancements in alignment since RLHF.
- ✓Roko
I think we should expect to die within a single digit number of years with ~80% probability.
- ✓Eli Lifland
If they don’t slow down, AI takeover is ~67%.
- ✓Abdul El-Sayed
If we don’t act quickly, we may get to a point at which we cannot act because it’s too late.
Disagree 6 people
- ✗Alex Karp
If we didn't have adversaries, I would be very in favor of pausing this technology completely, but we do.
- ✗Donald Trump
any attempts to pace the development of AI would give China the upper hand in developing the technology
- ✗China's Foreign Ministry
called AI companies' push for a slowdown "fear mongering,"
- ✗Andrew Feldman
Cerebras Systems CEO and co-founder Andrew Feldman thinks the AI doomer narrative isn’t doing the industry any favors.
- ✗Divyansh Kaushik
a pause right now would probably throw the economy into a bad recession and is not the right call.
- ✗Dean W. Ball
if we confidently knew that interpretability was on its way to being solved, I myself would be far more on the “full steam ahead” side than I would the “we should be careful” persuasion.
No clear side 6 people
- ~Fortune
it seems unlikely that these companies will come together to agree on a coordinated pause on AI development
- ~Dion Hinchcliffe
genuinely terrified them,” noting that previous public containment incidents have already been quite disturbing, and that there may have been private evaluations that showed even worse dangers.
- ~Dario Amodei
toughest dilemma
- ~AndrewCurran_
argues frontier “pacing” is now broken. One US lab defecting forces the others to accelerate
- ~Andrew Feldman
We have to be thoughtful about it. We have to regulate it reasonably,”
- ~Redwood Research
So we are quite uncertain about whether it’s good to help improve frontier monitors, and if it’s good whether it’s a better use of time than other research directions.
Side unclear 5 people
- ·Center for Humane Technology
We’re not in a race for the technology. We’re in a race for the technology times the rightly governed and wisely applied version of it
the big reason why Anthropic doesn’t want to pause is because they’re worried China is going to catch up.
- ·Theorem and Anthropic
multiple well-resourced labs at or near the frontier, in multiple countries, agree to stop under the same conditions.
- ·Dion Hinchcliffe
having a slower frontier with government oversight in the form of licensing, compute controls, export restrictions and so on strongly favors the incumbent model makers.
- ·teortaxesTex
He argues OpenAI and Anthropic are far enough ahead to “pace” between themselves
- ·A colleague
slowing down does lower your error rate, but mostly because you're doing less of everything
Week by week
- · 2: Import AI ✓, Zvi Mowshowitz ✓
- · 15: teortaxesTex ·, AndrewCurran_ ~, Andrew Feldman ✗, roon ✓, Jakub Pachocki ✓, Zvi Mowshowitz ✓, Roko ✓, Eli Lifland ✓, Divyansh Kaushik ✗, President Halla Tómasdóttir ✓, Center for Humane Technology ·, Dean W. Ball ✗, Abdul El-Sayed ✓, Redwood Research ~, A colleague ·
- · 9: Dario Amodei ✓, Sam Altman ✓, Elon Musk ✓, Jakub Pachocki ✓, Dion Hinchcliffe ~, Alex Karp ✗, Donald Trump ✗, China's Foreign Ministry ✗, Thomas Larsen ✓
- · 1: Anthropic spokesperson ✓
- · 1: Bernie Sanders ✓
- · 2: Theorem and Anthropic ·, Fortune ~
Built 2026-11-21 06:50 UTC from the news knowledge
graph by code. Every quote was checked against the source's own words; which side a quote takes is a model's reading,
made twice.