Arvind Jain explains which analyst roles AI will eliminate. Task-based data analysts who build dashboards on request without business thinking are the most vulnerable, while those who reason about business problems are not.

Enterprise AI token spend follows a steep power law. Some Glean users spend $10,000 to $15,000 per month while most spend around $20.

70% of Higgsfield's $300M ARR comes from agencies, not individual creators.

LTX Video's variable tokens architecture allows models to allocate more tokens to challenging areas of a scene such as physics.

Osirion has collected the largest ever olfactory dataset for AI training, comprising 5.43 million sniffs.

AI token spend among enterprise users follows a power law distribution, with some users spending $10,000-$15,000 per month while most spend around $20.

Grok 4.5 costs one-tenth of Fable's cost for equivalent tasks.

Jeff Currie explains how steep futures backwardation can generate large rolling returns even when spot prices fall, using oil as the concrete example.

'Super clipping' describes a frustrating power loss experienced by drivers when energy output is diverted to charge the battery instead of propelling the car.

Conductor introduced a pattern where a user-submitted prompt triggers an AI agent to implement a feature request directly, a workflow likely to spread to other products.

“The moon's gravity on Earth is the same no matter what the phase of the moon is.”

“40 to 41% of employees saying they ship AI work that they couldn't explain if asked. This is bot f***.”

“If you live in a state where there is an arbitrary power who can have you put to death, you are a slave.”

“Consensus 2027 EPS change for AI infrastructure stocks. Insane. 42%.”

“If I had a death wish, I would have just gone and done it. I would have spent nine years training for it.”

“You're spending a trillion dollars a year and the revenue on AI talk about maybe 70 80 billion.”