20 Photos and videos
Pinned Tweet
Jun 10
Any plan for surviving superintelligent AI that doesn't go through strong international coordination fails in at least one of three ways:
- It sparks war between nuclear powers
- It causes a misaligned ASI to kill everyone
- It establishes a permanent dystopian dictatorship
2
6
44
6,037
Evergreen.
cognition.cafe/p/the-realpol…
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
1
2
9
258
Alex Amadori retweeted
Updating the classics
24
216
3,426
88,053
Jun 11
This does not address catastrophic risks, and fails all three checks for a plan to address catastrophic risks.
Development, not deployment, of powerful AI needs to be restricted at a global level if we are to survive ASI.
x.com/DarioAmodei/status/206…
Jun 10
Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: darioamodei.com/post/policy-…
1
1
13
652
Jun 11
...be in a situation where everyone has to cut corners on safety to win the race, and we likely end up with an ASI killing everyone.
And whoever is losing still has a reason to attack preemptively with all their military might before the coalition gets a decisive advantage.
1
4
40
If you are still confused about this, I strongly recommend reading this article, written by @andreamiotti
and myself: "Anthropic did not call for a pause."
controlai.news/p/anthropic-d…
Jun 10
In contrast with the last Anthropic blog post, Dario's new one is back to softpedaling: five big subsections about "positive impact" and "securing leadership by democracies", with one throwaway line on "loss of control of AI systems" buried deep.
1
4
8
830
Jun 10
Any plan for surviving superintelligent AI that doesn't go through strong international coordination fails in at least one of three ways:
- It sparks war between nuclear powers
- It causes a misaligned ASI to kill everyone
- It establishes a permanent dystopian dictatorship
2
6
44
6,037
Jun 8
As a general rule, any capability that seems cool and powerful will accelerate death from RSI more than it accelerates any other theory of change.
This is because capabilities are much easier to specify than complex properties like "alignment" or "helps people with epistemics"
Jun 8
This will almost certainly help destroy the world by improving ASI-relevant capabilities more than it helps save the world by improving "epistemics".
Like seriously, you know like I know that one of the main bottlenecks in AI capabilities is memory-management for long-horizon tasks.
For any tool like this that you add to the AI toolbelt, it helps ASI capabilities much more, because those can be improved with a for-loop that uses numerical metrics from the world as a reward signal.
Things like "alignment" or "epistemics" will go much slower, because you can't improve them in a for-loop.
You either a very tasteful human in the loop, or "all of society" digesting it for a long time.
3
30
Alex Amadori retweeted
Jun 5
Not just foreseeable, but foreseen and called out.
We ran a campaign to get them out of the first ever AI Safety Summit. We won that one.
Yet most of the field kept licking the boot of the companies and here we are.
controlai.com/past-campaigns…
Jun 4
What a waste of a few years RSPs were. Unfortunately the waste was foreseeable.
1
15
51
2,950
Alex Amadori retweeted
Jun 3
🚨NEW: We’ve just launched our campaign in Canada!
A cross-party coalition of over 30 MPs and Senators are calling for Canada to negotiate an international prohibition on the development of superintelligence, recognizing the risk of human extinction posed by the technology.
🧵
14
57
132
32,737
Alex Amadori retweeted
May 22
UK MPs were just selected to introduce bills of their choice.
And we've published the updated version of our bill to ban superintelligence we presented to Number 10!
If introduced, it'd be a world-first.
Ask them to introduce it using our contact tool: ukbill.controlai.com
3
18
38
3,587
Alex Amadori retweeted
May 5
While AI leaders warn superintelligent AI could cause human extinction, governments are asleep at the wheel.
The companies building superintelligence admit the danger, yet expect to create the tech within a few years.
My piece in @spectator on the threat and what we can do.
14
30
79
6,110
Apr 30
There are two wolves inside of Opus: one of them agrees with the user, the other one also agrees with the user.
25
Alex Amadori retweeted
Apr 29
We at @ControlAI are sometimes asked what we think of banning datacenter construction.
At CAI, we focus on one issue: The risk of extinction from superintelligent AI.
The only way to prevent this is to prevent the creation of ASI.
Datacenter bans do not help with this goal.
19
23
103
8,900
Alex Amadori retweeted
Apr 21
We have a plan to turn $50M / year into a 10% chance of banning ASI.
An international ban on superintelligence doesn't happen unless the public and policymakers understand extinction risk from superintelligent AI.
Our job is to inform them. If no one does it, it won't happen.
2
10
34
1,771
Alex Amadori retweeted
Apr 21
We think ControlAI can turn $50M / year into a 10% chance of banning ASI.
Most of the AI safety community has been far too coy about extinction risk. We're not.
It's not that complicated: AI smarter than humanity poses an unacceptable risk of human extinction, and we need an international ban on its development.
Here's our plan to get this done.
34
25
223
106,542
Apr 20
insta reels algorithm when it makes you scroll for 5 hours straight
42
Alex Amadori retweeted
Apr 13
My horrendous nightmare of a political lifecycle, ladies and gentlemen and others.
6
10
292
11,074