Ash Jogalekar
@curiouswavefn
Followers
29K
Following
53K
Media
9K
Statuses
54K
History of science, military and U.S. history, physics, nuclear energy, nonproliferation, biotechnology. Books. Scientist in Residence at Oppenheimer Project.
San Francisco Bay Area
Joined May 2009
1/n: Very pleased to see this piece which I worked on for a long time out in Fast Company. In it I apply "mosaic theory" to LLMs, pointing out how they can unintentionally assembly benign information into a questionable whole. A summary 🧵
fastcompany.com
Like a student who once designed a nuclear bomb from textbooks, today’s AI systems can stitch together public scraps of information into dangerous blueprints—at speed, at scale, and without realizing...
5
15
57
3
2
47
Stake your $SOL with Everstake — trusted by 1,000,000+ users. Compatible with all major wallets: Phantom, Solflare, Exodus, Ledger. A Top 10 Solana validator with 6.2M SOL delegated. • 6–7% APR • Low 7% fee • Proven reliability
0
0
9
Well said. Unfortunately social media and our economy these days reward trendy, fleeting, buzzwordy knowledge rather than fundamentals. What's ironic is that the trendy, fleeting, buzzwordy knowledge wouldn't exist without the fundamentals.
Dear “15-18 yo founder”s sending me DMs, don’t. Go and hug your parents, fall in love, eat chocolate cereal for breakfast, read poetry. Nobody will give you back these years. And sure, do your homework and learn math and code if that feels fun. But stop building SaaS and
0
3
16
A lot of these stories are making the case for the extraordinary and unprecedented advantages immigration has bestowed on the U.S., a case that especially needs to be made right now.
Left Bulgaria at age of 15 to go to colleague in the US studying physics. Saw my family 4 times in the span of 20 years. Worked every Christmas and other holidays. All my grandparents died without me knowing. I haven’t met most of my cousins. My mom died without me knowing and
1
0
10
Some interesting sightings yesterday, including two from the horses’ mouths.
2
28
336
Today’s acquisitions. The Chandrasekhar volume belonged to physicist Bruno Zumino.
1
2
67
Great Edison anecdote demonstrating that you should use math only when you need to.
17
62
255
Why was GPT-6 afraid of GPT-7? Because 7 8 9. You're welcome.
GPT-6 will be renamed GPT-6-7, you're welcome
1
2
11
@yacineMTB if i were like, a sports star or an artist or something, and just really cared about doing a great job at my thing, and was up at 5 am practicing free throws or whatever, that would seem pretty normal right? the first part of openai was unbelievably fun; we did what i believe is
1K
863
17K
Night Thoughts on the Promise of Agentic AI in Drug Discovery https://t.co/IYRgr4gzVT
0
0
5
Paul Dirac: In the absence of experimental evidence, basic beliefs of theoretical physicists may initially have almost a religious flavor, guided by faith and aesthetics. Fortunately unlike religion, these beliefs soon face the hard test of experiment.
13
21
193
physicists are very antisocial by nature. if one ever tells you they're interested in forming a group, you can be sure it's a Lie.
47
179
1K
I understand the sentiment, but we might be confusing means with ends. DeepSeek showed the ability to build very good LLMs using inferior chips.
I would simply not give away the essential bottleneck input for the most important dual-use technology of our era to the US's primary geopolitical rival.
0
0
3
I find the “tsunami powerful enough to devastate a coastal city” claim *extremely” implausible. The mechanisms of an earthquake-tsunami and a nuclear explosion, even a very big one, and the energy dissipation are very different. https://t.co/LNzG4uvg1c
nytimes.com
The test comes days after President Vladimir V. Putin announced the previous trial of another Russian nuclear weapon.
1
0
4
@curiouswavefn A truly great book for those starting out. I discovered it by accident in the library on a work term after my first year and it had an enormous influence on me.
0
1
9
"We are human beings first and scientists second, because knowledge implies responsibility." - Freeman Dyson
This is very good. Even if you disagree with the conclusion, much of the argument is correct. In particular, OpenAI's pretence that they're not moral actors is obviously false. Disavowing any responsibility by saying "it's the user's choice" is also an argument used by people
0
1
13
List of books Dyson bought from school prize money (15 shillings) when he was a student at Winchester College. 60 years later he still had them on his shelf.
5
31
273