Loading activity…
@evanjconrad
ceo @sfcompute the american culture is that we start from scratch and we beat you
AI
A post quotes a client interested in partnering with a known AI lab to develop the mall property.
AI
A user expects a small amount of inference to retain fixed pricing, but at very high prices.
AI
A commenter argues an AI data center would not be viable there and blames journalists for misleading the public.
AI
The user says their company's token bill appears to keep rising and argues one of those figures must change.
AI
A post quotes Apollo's chief economist on diverging growth forecasts; another argues early-market demand is often underestimated.
AI
A user quote-posted another message, reacting with just one word: 'incredible.'
AI
An essay argues that falling prices and surging token use favor platforms that control distribution over benchmark-leading labs.
AI
SF Compute says each cluster node has eight H100 GPUs and bills by the minute once ready.
AI
A user claimed there was a GPU shortage and joked that a 9% MFU should prompt government intervention.
AI
A quoted blog passage estimates that nonstop agents could supply as many weekly working hours as about 140–720 million full-time employees.
AI
The poster also lists IB and cluster-local object storage and claims an unusual buying arrangement makes buyers money.
AI
A commentator argues that tech journalists too often treat a speaker’s potential ulterior motives as evidence that their claims are false, especially in AI debates.
AI
A user argues that companies raising venture capital for compute deals rather than taking on debt may lack confidence in pricing.
AI
SemiAnalysis reports local moves to pause data centers, alongside a discussion of emerging oversupply in early-stage projects and why positive views of AI don’t always extend to the facilities.
AI
One post argues that failing to grasp AI's dangers leads to bad investment decisions. Another imagines explaining missed superintelligence profits to a fund's investors.
AI
The company says it builds and manages large-scale supercomputers, and frames the role around earning the trust placed in it.
AI
Amodei says Anthropic will give third-party evaluators permanent, employee-level access to its systems to verify adherence to safety measures, report incidents and assess models’ alignment during training.
AI
A team member says San Francisco Compute backs the system and will let users scale up without reservations as its large-scale clusters come online.
AI
The builder also describes reinforcement-learning sandboxes that can save and restore disk and memory state, claiming roughly 7x savings by freezing them during tool calls.
AI
The post invites collaboration on what it calls “the real bottlenecks in compute.”
