
Senior writer and editor @coeff_giving. Author of END TIMES. Email [email protected]
New York
Joined October 2009
- Tweets3.6K
- Following2K
- Followers20.2K
- Likes1.2K
Pinned Tweet
Some professional news: after four and a half years at Vox, I’ll be joining @coeff_giving as a senior writer and editor, with a focus on global catastrophic risks.
"4. Youth unemployment in China has gone from non-existent to one of the highest in the world during a period when its poverty rate has plunged. Weird."
That last part I continue to find so strange...
Young people around the world are not okay. Incredible graphs from WSJ
1. Unemployment rates for young college grads have been climbing compared with the broader population for years
2. AI might be exacerbating the employment challenges for young ppl in software
3. In Italy and Spain, more than 80% people in their 20s live with their parents (!?!!?)
4. Youth unemployment in China has gone from non-existent to one of the highest in the world during a period when its poverty rate has plunged. Weird.
129
"The American people do not want to pay for your political project" theargumentmag.com/p/how-to-…
What must a $1 million housing grant deliver to pencil out @coeff_giving? About 2,100 additional homes it can fairly claim, by @aarmlovi’s math for Abundance and Growth Fund. That’s after discounting for uncertainty, what would’ve happened anyway and everyone else’s work abundanceandgrowth.org/p/hou…
One year ago Open Philanthropy became @coeff_giving, and I definitely think by the next anniversary I'll remember to stop using the old name coefficientgiving.org/resear…
Bryan Walsh retweeted
We’re releasing a concrete proposal for how a near term AI slowdown could work. It involves:
1. AI companies declaring how they are using their compute and letting third party auditors verify those disclosures. We call this Verified Compute Allocation Disclosures (VCAD).
2. Compute Reallocation towards serving customers and safety, away from capabilities R&D.
(2) slows AI capabilities , while (1) makes the slowdown credible.
AI companies currently seem to be going for ‘safety case pacing’. But: (i) there’s no consensus on what a valid safety case looks like so there’s no way to consistently audit for compliance, and (ii) if the R&D doesn’t get paused, temporarily delaying a training run for safety doesn’t meaningfully slow the overall trajectory of capabilities progress. It would be great if we had a viable safety case regime, but we need much more time to get it working.
Compute reallocation is a good pacing proposal because it is relatively specific, can be verified, and would lead to a more durable slowdown.
I find this remarkable - in a good way - given how much our views of crime tend not to match reality news.gallup.com/poll/716657/…
162
Hmm, but what if what the AI models WANT us to do…
NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say
AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems
washingtonsun.com/technology…
378
Bryan Walsh retweeted
Will economic growth from AI reach the developing world?
We think through possibilities and what they mean for policy and research: new paper with @LagakosDavid and @MitchVanVuren. 🧵
A far more boring but effective robot tax
Tax AI*
* end the deductibility of interest expenses
taxpolicycenter.org/taxvox/t…
Bryan Walsh retweeted
Thought this report from @Research_FRI was interesting (and covered a lot more than I got from the tweet thread) forecastingresearch.org/rese…
AI companies increasingly say they may soon reach recursive self-improvement (RSI).
With RSI, AI capability progress could accelerate dramatically. This acceleration could happen inside AI companies before outsiders can notice.
How close is RSI? Our new report reviews the evidence, turns key disagreements into forecasting questions, and gives initial LLM forecasts. 🧵
Bryan Walsh retweeted
At some point, all these network failures with NO evidence of a cyberattack seems almost worse than finding evidence as “undetectable cyberattack” becomes more likely
News: The highly classified network that the National Security Agency uses for its internal communications is down, five people familiar with the matter tell @NatashaBertrand, @KatieBoLillis, @snlyngaas & me.
The network, called NSA Net, allows employees to communicate classified information internally. At this point there is no evidence of a cyberattack, sources said.
As the agency works to understand the scope of the problem, it has enacted staggered arrivals for personnel, the sources said.
A cyber autoimmune disorder?
There are lots of outages... Why? AI attacks?
My more normal guess is that because there have been a lot of AI-discovered cyber vulnerabilities (including lots probably not yet disclosed), companies across all industries are having to do a lot of patches and upgrades, probably faster than usual. Such upgrades can sometimes go wrong, causing outages.
184
Who among us hasn't tried to play the same card after getting dumped?
Breaking News: Jensen Huang said he would return as Nvidia’s chief executive if asked, his first public statement since resigning. He resigned last week after revelations that he had affairs with an employee and a woman linked to Chinese intelligence. nyti.ms/4iL8g5J
253
Good news for the media business: Americans spend 91 hours a week with media — 6 more than they spent in 2018. Bad news for the media: all of that increase comes in background attention, when we're doing something else. Focused attention fell by an hour: bcg.com/publications/2026/vi…
I think I knew this but I still find the chart striking ourworldindata.org/data-insi…
135
A lot resonated for me with @KelseyTuoc piece on the existential challenges of a different kind we'll face if we're not destroyed by the more directly existential challenges of superintelligence. A base level fear for her is that she would not recognize society after ASI. I share that. But I can't help wondering as well — how much would our ancestors recognize ourselves in the mostly secular technological world we have now, the kind she celebrates in the passage below? Would they not also feel that — despite our success banishing suffering that was so common in the past — that we had lost something essential? theargumentmag.com/p/what-ha…
162
Short timelines are significantly shorter than movie timelines. For that matter, long timelines might be too
Bryan Walsh retweeted
Earlier this year, @nanransohoff wrote about a coming AI-driven "third wave" of American philanthropy. Since then, I’ve had some conversations where people worried there won’t be enough good places to direct that $.
But I think the new wave could be spent amazingly well, many times over. In a new piece, my colleagues @Chris_topian, Aaron Roesch, and I catalog many extremely impactful ways you could spend the third wave, from ambitious & urgent AI safety plays to improving the conditions of farmed animals to closing the extreme poverty gap: coefficientgiving.org/resear…
If Nan's base estimate of the entire third wave (~$37B/yr) went to cost-effective global health programs, it could save something like 2 million additional lives per year - I think that is a wild, wild amount of potential positive impact. And the potential impact isn't restricted to global health.

























