On the very first day of his second term, President Trump revoked Biden’s AI executive order, which required AI companies to disclose details of their internal safety testing.
It says a hell of a lot that he's changing course. So what gives?
https://t.co/aW1n5l5auQ
It wasn't very long ago at all that JD Vance and Ted Cruz were happy to mockingly dismiss AI safety and regulation.
But since Trump's decided to take it seriously, they'll have to take it seriously too.
So what's led Trump to this surprising point?
https://t.co/aW1n5l5auQ
(By the way, if you're interested in where AI money's moving in the midterms, might we recommend again that you have a little look at our AI campaign finance tracker?) https://t.co/ntwQkOjx9Z
In our ongoing coverage of how AI money's moving in American political scene: is AI safety’s ‘hard money’ its secret weapon in the midterms?
Let us explain:
https://t.co/6nlS2RyVtv
AI is taking over such jobs as beating Pokemon Red, winning at least one short story competition, and creating ironicaly fake quotes for a book about AI's effect on truth.
Keep up with everything else in AI this week via our delightful newsletter:
https://t.co/544gvy3uky
Just in case it's handy, you can keep up with the rapidly moving and terrifyingly unknowable bleeding edge/jagged frontier of AI with our newsletter (e.g. here's some some pretty interesting/unnerving stuff from this week's).
Why not have a look?
https://t.co/544gvy3uky
Our writer Celia Ford reports from the AI safety conference ControlConf, where a peculiar roleplaying game elucidated some quite important, surprising and frightening ideas around AI alignment v. AI control...
Just three hours before Trump was set to sign an executive order that would've established pre-deployment evaluations for frontier AI models, he pulled the plug.
This isn't really an ideal way to run things if you take AI at all seriously.
If you happened to be wondering how to beat a tech giant that wants to build a data centre in your area (and you have a particularly good reason to), a retired tech exec seems to have figured out the trick...
https://t.co/qmM545LE2L
One of our intrepid writers recently larped as a rogue AI at an AI control conference — and learned a surprising amount about the limits of AI control.
If you're curious about something strange and frightening about the near-future, click through here:
https://t.co/Zi9Jt8Btn1
It might be worth taking a step back from the story we've been given about the great US-China AI race and ask if it's actually true — and if it's not, who exactly this story serves, and if we might be able to tell a better one.
https://t.co/Nmr9D2mMob
China doesn't seem to be in the same race to a existentially-risky superintelligence that the US thinks it's in — and the idea it is might make that existential risk all the existentially riskier (for all of our existences).
https://t.co/Nmr9D2mMob
It's another Transformer weekly newsletter!
From $900b (yes, $900b) valuations to the "sad wives of AI" losing their husbands to chatbots, it's put everything interesting in AI lately in one handy place 👍️
Just click through and have a look around:
https://t.co/WuAWv5oZOu
Here's a hot take: Palantir's hot takes are just a cheap way to manipulate the attention economy so as to manipulate the actual economy.
Here's some clickbait: we break down how that all works in the link below.
https://t.co/nEwzW8VKYY
Some people might wonder why the data-gathering company Palantir named themselves after the evil, all-seeing stone from The Lord of the Rings that gradually corrupts anyone who uses to use it.
It seems it might just be that the controversy is the point.
https://t.co/nEwzW8VKYY
We've been sold the story of a tense AI race between the US and China with apocalyptic stakes — but it might not be true, it might be serving particular interests, and it might be leaving the rest of us to deal with some pretty severe consequences.
https://t.co/Nmr9D2mMob
Earlier this week, more than 600 Google employees signed a letter to their CEO Sundar Pichai protesting the company’s ongoing negotiations with the Department of Defense — only to find out at the same time as everyone else that it had no effect at all.
https://t.co/Pf0vU6hO2D
The Transformer newsletter is here, with everything you'd want to know about AI across the last week, such as this extremely conveniently bullet-pointed summary of what's been going on with Elon Musk's $134b lawsuit against OpenAI. So now you know.
https://t.co/YvVuDRvJ6h
The fine print matters:
“’Should not be used for’ is not the same as ’shall not’ or ’will not’ be used for. ’Should not’ imposes no enforceable obligation on the Pentagon.”
Some Google employees are speaking out against their new Pentagon deal:
https://t.co/Pf0vU6hO2D