Apply-focused indie lab researching mathematical foundations of systems across domains. Ex-founder (fintech/infra), CS + big data/distributed systems (Adobe).
Mill Valley, CA
Joined April 2007
- Tweets3.2K
- Following1.8K
- Followers692
- Likes1.5K
💾saved today: "Every new idea in science must pass through three phases. P1 everybody says the man is crazy and the idea is all wrong; P2 they say that the idea is correct but of no importance; P3 they say that it is correct and important, but we knew it all along." - Freud
2
68
Microsoft, Amazon, Google guarding the private corporate data sheepfold
theregister.com/2026/03/26/g…
1
24
What will every pumped SaaS that can't monetize users any longer do?
"Postman silently auto-installs AI plugins and attempts to share all API data"
news.ycombinator.com/item?id…
github.com/postmanlabs/postm…
1
1
166
ℂosmin Lehene retweeted
cryptocurrencies should stay in their lane rather than try taking over what banks are already good at
As intellectual property (and thinking in general) value goes to zero, things like energy, compute, connectivity need to become human rights.
1
18
Datacenter capacity is already getting sold out datacenterdynamics.com/en/ne…
1
25
you don't need to shift from personal computer to the other extreme of datacenter to increase utilization.
Fine grained resource sharing enables fine control over the tradeoff of resource utilization and resource access.
1
23
better utilization ≠ fair allocation, fair pricing
the only chance to recoup wasted dollars is to enforce a (soon unnecessary) middleman by monopolizing computation power. Out-pricing consumers is one way, pushing policy against open or local models is another
OpenAI CEO says it's 'quite terrible' for your PC to have RAM — Sam Altman says data centers offer better utilization tomshardware.com/tech-indust…
1
71
inability to interpret your models is inability to interpret the cost/benefit tradeoffs from sharing their outputs
1
64
Sam be like: drop Russian. just roulette. it's cleaner
OpenAI's longtime safety lead David Robinson has quit Friday and now he wrote a strongly negative piece on The Atlantic.
> "Companies do not have anything close to certainty that good scores on their alignment tests actually mean a good model"
> "there is a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term."
> "Today and tomorrow’s AI systems are far more capable and dangerous than the systems we were building even six months ago."
64
