
The DeepSeek 4 Pro GA is a very good model. But holy sh*t, the pricing is so damn good! $0.435/M input and $0.87/M output @ 1M-token context This is insane pricing!!

By
–

The DeepSeek 4 Pro GA is a very good model. But holy sh*t, the pricing is so damn good! $0.435/M input and $0.87/M output @ 1M-token context This is insane pricing!!

By
–

You know whats crazy? Grok 4.6 is not only cheaper than Opus 5 and 5.6 Sol. Its even cheaper than Sonnet 5. While at the same time, at least on par with the sota models. *Thats* the real moat.

By
–

.
@BharatKChandar
, @RuyuChen and I just released an updated version of our paper "Canaries in the Coal Mine?" The headline is that we still don't see any widespread job displacement due to AI, but the earlier trends we identified hold up or even grow. E.g., the relative decline
By
–
“Pay attention: This is where the A.I. build out goes from something that will threaten tech companies to something that will threaten our entire banking system and our entire economy. Banks, ever ready to package dodgy debt for other people to hold, have begun operation
By
–
Exactly. Access to the same models does not guarantee the same outcome, because infrastructure, local capabilities, and industrial execution determine how much economic value stays in the country.
By
–
A purpose-built model can run 10 to 100x more cheaply than a frontier model on the narrow task it was trained for, with the weights owned by users. There, it is possible to run data synthesis, evals, model training, and deployments. Test it out

By
–
CFO, CEO, financial analyst, tech entrepreneur, student, job-seeker, friends & family — be future-ready with this book >> "Quantum Machine Learning and Optimization in Finance" (494 pages; 2nd Edition): http://
amzn.to/4lNlBt5
By
–
the wild part is every runtime on that list is free to licence. you’re paying $14M for machines doing nothing between turns

By
–

THE SAME BILL CAN BE $14M OR $44K One million agents running for one month. Same work. Four ways to pay for it. → One VM per customer on AWS: ~$14M
→ OpenClaw / Paperclip: ~$1.5M
→ Hermes / Eve: ~$740K
→ Naive serverless: ~$44K Now the part [Traduit de FR vers EN]
By
–
→ A chatbot reply can cost about 4 cents. An agentic task can run roughly 100x higher. → 1,000 agent tasks a day can mean close to $1.3M a year in cloud spend. → Run your own AI factory on @HPE infrastructure and @NVIDIA compute, and the marginal cost per token nears zero.