I mean, that's why the effort can be changed to lower settings so you don't slow down via excessive token usages
ECONOMY
-

A Reported 2.2% of US Households Pay for AI Subscriptions
By
–
Wild to think only 2.2% of U.S. households have a paid AI subscription. The AI revolution has barely even started.
-
Open Weights Could Shift AI Value Toward Compute Providers
By
–
I am confused about the belief that if open weights eventually dominate it will lead to the collapse of AI. If the Labs lose (which is not happening now), it isn’t because AI was useless: compute is still the barrier & compute providers will capture the value rather than Labs.
-
More Compute and Less Regulation May Not Restore AI Leadership
By
–
Can the US GenAI industry be saved? How? @davidsacks has called for more data centers and less regulation but it’s not like we have much substantive regulation now (and we already have more compute) and yet our lead has diminished anyway. Anyone got a better idea?
-

Lower-Cost Inference Helps Specialized Agent Reinforcement Learning
By
–
Agents are becoming specialized, using reinforcement learning post-training to develop domain-specific skills Every post-training rollout is an inference call, so a lower cost per token flows straight into a higher Intelligence per Dollar on every run. That is what the
-
Maximum Reasoning Effort Is Not Always the Best Model Setting
By
–
Yes, but it's valuable to explore how to use the models the best way, instead of just yoloing everything on max.
-
Subsidized Cloud Tokens May Shape the Case for On-Premises AI
By
–
Eventually, subsidized tokens will dry-just like Uber free rides, DoorDash, etc-and these people will be left with no option but moving to on-premises. Only, it will be extremely more difficult to do by then… That's why I tell everyone they should start today.
-
A Chip-Backed Loan Finances Inference Silicon Beyond GPUs
By
–
Congrats to @fastinference on landing $400M from Upper90 — the first chip-backed loan built around inference silicon, not GPUs. This is exactly the shift we've been building for: fast, efficient, power-smart inference that doesn't need water-cooled supercomputers. Read the
-
Affordable Electricity Shapes Sovereign AI Compute Capacity
By
–
Just as cheap electricity determines the industrial powers of the past, it will now determine the AI superpowers of the future.
— Nina Schick (@NinaDSchick) 17 juillet 2026
Sovereignty in this century is found in the physical ability to process Intelligence at industrial scale. pic.twitter.com/kt5Yz3HMJBJust as cheap electricity determines the industrial powers of the past, it will now determine the AI superpowers of the future. Sovereignty in this century is found in the physical ability to process Intelligence at industrial scale.
-

Databricks Links Growth to Enterprise AI Gateway Adoption
By
–
Databricks CEO @alighodsi joined @CNBC
's 'The Exchange' to break down the company's new $188B valuation, and what's fueling accelerating revenue across regions and products. That includes more companies adopting tools like Unity AI Gateway to manage spend and access across