Can AI and Energy Scale Together? AI is no longer only a software story. It is becoming an infrastructure story, one shaped by electricity, grids, data centers, cooling, connectivity, capital, and long-term energy planning. That is why the convergence of artificial
ECONOMY
-
Transformative AI Economics Needs Guardrails, Incentives, and Institutions
By
–
We'd like to draw attention to the need to do more to understand the economics of transformative AI and to build the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society. I'm glad you've been helping with the
-

Changing the Agent Harness Cuts Task Cost and Runtime in Study
By
–

2/ The Harness Effect A study ran 22 tasks across 6 foundation models, then changed only the orchestration harness and held the models fixed. The harness alone cut cost per task 41% and median wall clock 44%, with tokens down a similar amount and completion quality at parity.
-
Google’s AI Strategy Differs From OpenAI’s and Anthropic’s Revenue Pressures
By
–
I think people are underestimating Google. OpenAI and Anthropic are pure AI companies. They need to get to profitability so they put a lot of focus on consumer and enterprise solutions to make money. They have to keep shipping new models and harnesses to keep people using them.
-

Writing Benchmark Finds Terra Performs Better With Reasoning Turned Off
By
–
GPT-5.6 Terra’s default setting is its worst one for writing with reasoning off, “terra none” beats “terra medium” at the same price, and nearly matches “sol” at 40% of the cost, meanwhile “fable 5 xhigh” matches “fable 5 max” exactly, at half the price the only model where
-

Fable XHigh Matches Max on a Writing Benchmark at Half Cost
By
–
You may be wasting tokens with Claude Fable 5 max effort for our writing benchmark, it’s just wasting tokens. at half the cost, “fable 5 xhigh” exactly matches “fable 5 max” (89.0 vs 89.0), while “sol” peaks at high and gets worse at xhigh so for voice & tone, and writing
-
Access to Frontier Models Shapes Vendors’ AI Strategy Arguments
By
–
Lots of the AI & corporate strategy talk on X is not that complicated: vendors without frontier models are always explaining why you shouldn't trust the frontier model companies. Vendors w/frontier models are selling solutions based around frontier models. (Either could be right)
-
Perplexity Chooses Grok 4.5 for Evaluations, Cost, and Zero Retention
By
–
Two reasons why we integrated Grok 4.5 inside Perplexity Computer within a few hours: 1) It scored the best on our evals and was the most cost effective option 2) ZDR was available from the get go and that’s what our customers want
-

Economists and AI Researchers Warn of Rapid Economic Change
By
–
16 Nobel Prize winners and over 200 economists and AI researchers signed "We Must Act Now", a statement organized by Stanford's Digital Economy Lab. The warning: AI could bring economic change "larger than the Industrial Revolution, but unfolding over a vastly shorter time
-

AI Research Raises Prospects for Faster Economic and Scientific Progress
By
–

The most important AI researchers of our time take AI seriously – not only because of its capabilities, but because it could accelerate economic growth and scientific progress at an unprecedented pace. We should take it seriously, too.
