LLMs have advanced much further in fundamental mathematics than in other foundational fields. However, pushing the frontiers of science and technology will depend largely on breakthroughs in fundamental physics and biology. Progress in these two fields is often constrained by the
Looking at predictions like these, I think Demis Hassabis was absolutely right to pursue AI applications in science and mathematics, but he moved too early. He should have focused more on improving AI’s fundamental capabilities.
Another thought comes to mind: if AlphaFold was
FUNDA Platform has observed that Hopper prices are starting to decline, while Blackwell remains very strong.
FUNDA explains the reasons behind this divergence in its latest AIDC Report:
This isn’t a demand rotation from Hopper to Blackwell. It’s a supply-tightness gap between
Deep|AIDC: 2027 Demand Runs at Roughly 2x Deliverable Supply, and the Binding Constraint Is Power
North American AI data center demand in 2027 runs to roughly 35GW on an IT basis, against 16.5-23.4GW that can actually be delivered. Where our two independent supply methods
For a major technology company such as Anthropic, would it make sense to publicly file its S-1 during Labor Day week? After the S-1 is made public, how long does it typically take for the company to formally list, and what does the process involve?
Detailed Repot
September 7 is Labor Day. Especially after the grueling July everyone went through this year, many people may want to take a longer break. So filing the S-1 in late September would be perfectly reasonable. A listing would normally take place three to four weeks after the S-1
At the same time, GPT-6 is likely to be the first model to incorporate latent thinking, which would make distillation significantly more difficult. Fable 5.1 has also made it considerably harder to transfer signatural data from larger models to smaller ones, further raising the
