Leary

10 ms·

1.
GPT-6 Astra Benchmarks Image(cdn.thenewstack.io)
9 points by Leary 14d ago|1 comments
2.
The Fable 5 Export Controls Harm US Cyber Defense(lutasecurity.com)
10 points by Leary 3mo ago|1 comments
3.
by Leary 3mo ago
Uploaded my code base and it forced switched to Opus 4.8 after thinking for 5 minutes even though I prompted it to not work on cybersecurity related things. Amazing.
4.
by Leary 4mo ago
news from 2025
5.
Anthropic "Gift Max" Exploit cost user €800, tanked SCHUFA score, and a ban(old.reddit.com)
3 points by Leary 5mo ago|0 comments
6.
Anthropic's 'anti-China' stance triggers exit of star AI researcher(yahoo.com)
3 points by Leary 11mo ago|1 comments
7.
by Leary 1y ago
“Large frontier developer”: a frontier developer that together with its affiliates collectively had annual gross revenues in excess of five hundred million dollars ($500,000,000) in the preceding calendar year. So any large company training
8.
by Leary 1y ago
Except that does make sense because the new country would be the weighted average of the average CO2 contribution for both Americans and Nigerians.
9.
by Leary 1y ago
Amazing analysis, except for the fact China was dirt poor before communism.
10.
by Leary 1y ago
No way. I read Noah Smith (number 1 economist on substack) every week and he says one must measure China in per capita terms for all the good stuff: GDP (preferably nominal), and in aggregate terms for all the bad stuff (pollution, carbon d
11.
by Leary 1y ago
But how are you gonna foresee the total annihilation of the Chinese semi industry without Jordan the necromancer interpreting party documents for you?
12.
Y Combinator Brief in Epic vs. Apple(scribd.com)
5 points by Leary 1y ago|0 comments
13.
by Leary 1y ago
Does anyone know which technology on this tree has the most descendents?
14.
by Leary 1y ago
https://metr.github.io/autonomy-evals-guide/gpt-5-report/
15.
by Leary 1y ago
METR of only 2 hours and 15 minutes. Fast takeoff less likely.
16.
by Leary 1y ago
GPQA Diamond: gpt-oss-120b: 80.1%, Qwen3-235B-A22B-Thinking-2507: 81.1% Humanity’s Last Exam: gpt-oss-120b (tools): 19.0%, gpt-oss-120b (no tools): 14.9%, Qwen3-235B-A22B-Thinking-2507: 18.2%
17.
by Leary 1y ago
Did it with https://sites.google.com/site/jiaxiongyao16/nighttime-lights... USA (2013-2023 CAGR: 2.3%) 2014: 6.2% 2015: -5.3% 2016: -1.8% 2017: 15.2% 2018: -4.9% 2019: 4.5% 2020: -5.4% 2021: 6.7% 2022: 14.5% 2023:
18.
X-Master as Foundation: Can We Lead on Humanity's Last Exam?(arxiv.org)
1 points by Leary 1y ago|0 comments
19.
by Leary 1y ago
0:10 into the video, that's light skinned to you?!
20.
Reddit sues AI startup Anthropic for breach of contract, 'unfair competition'(cnbc.com)
9 points by Leary 1y ago|0 comments
21.
FTX Dumped Cursor Stake for $200K – It's Now Worth $500M After $9B Valuation(msn.com)
1 points by Leary 1y ago|0 comments
22.
by Leary 1y ago
Translation: You provide the capital and the data, we'll co-own the data centers share the models until Trump and the US government decide to shut it off as a bargaining chip.
23.
by Leary 1y ago
They are not "English-first". Deepseek-R1, for example, reasons in Chinese when you ask it a question in Chinese.
24.
China accuses US spies of Asian Winter Games cyberattacks(france24.com)
5 points by Leary 1y ago|0 comments
25.
The Fiscal and Economic Effects of the Revised April 9 Tariffs(budgetlab.yale.edu)
1 points by Leary 1y ago|0 comments
26.
Trump's Tariff Formula Makes No Economic Sense. It's Also Based on an Error(aei.org)
256 points by Leary 1y ago|191 comments
27.
by Leary 1y ago
Nah, ByteDance will sell if the US lets Nvidia sell them some B200s
28.
Researchers suggest OpenAI trained AI models on paywalled O'Reilly books(techcrunch.com)
5 points by Leary 1y ago|0 comments
29.
What I'm Hearing in China This Week About Our Shared Future-Tom Friedman(nytimes.com)
4 points by Leary 1y ago|0 comments
30.
by Leary 2y ago
So OpenAI's ChatGPT output cannot be copyrighted and it's legal to distill it?

More ›