“The pricing moat that frontier labs have been sitting on just collapsed. You can now get frontier intelligence at commodity prices, and it's only going one direction from here” – exactly what i predicted in August 2023. The tech community shunned me. Now you know why.
ECONOMY
-
Comparing Kimi’s Per-Task Cost With Fable and GPT-5.6 Sol
By
–
Not sure what you mean but Kimi was cheaper per task! 5x cheaper than Fable, and similar to 5.6 Sol high specifically
-

China’s AI Progress Raises Questions About Capital Spending
By
–
China is catching up in AI despite significantly lower capital expenditure, while Europe continues to lag far behind. I looked at the numbers, and the conclusion is clear: despite spending around 90 percent less on capital expenditure, China is managing to catch up with Western
-

ACRouter Selects AI Models by Task to Reduce Costs
By
–
ACRouter picks the smartest #AI model per task, beating Opus-only setups by 2.6x on cost
by @bendee983 @VentureBeat Learn more: https://
bit.ly/4wyeSrW #ArtificialIntelligence #GenerativeAI #AIAgents #AgenticAI -
Enterprises Buy AI Compute Faster Than They Can Measure Its Cost
By
–
The AI compute gap: Enterprises are buying infrastructure faster than they can measure what it costs
-
China’s Investment and Talent Are Driving AI Progress
By
–
The real story is that China is a large country, with lots of people, many of whom are very capable, and is investing heavily in AI. So of course they're getting very good at it.
-
Reducing Agent Token Spend by Improving Search Results
By
–
Agents that search more, reason less.
— Akshay 🚀 (@akshay_pachaar) 17 juillet 2026
(+ a strategy to cut token bills by 76%)
Here's a pattern that shows up in almost every agent trace: the model runs a search, gets backlinks and thirty-word snippets, and then has to go read the web itself.
That means fetching the pages,… https://t.co/OLQ7vn9lDq pic.twitter.com/U9TToLLVHhAgents that search more, reason less. (+ a strategy to cut token bills by 76%) Here's a pattern that shows up in almost every agent trace: the model runs a search, gets backlinks and thirty-word snippets, and then has to go read the web itself. That means fetching the pages,
-

OpenAI, Meta and SpaceXAI Compete on AI Model Cost Efficiency
By
–
OpenAI, Meta, SpaceXAI Compete for More Cost-Efficient #AI Models
by Lorelei Smillie Rachel Metz @business Learn more: https://
bit.ly/4fbJed0 #GenerativeAI #LLM #ArtificialIntelligence #MachineLearning #ML -
Smaller Open Models Raise Questions About Frontier Infrastructure Spending
By
–
ELI5 why we should invest trillions in data centers and closed source LLMs when LLMs have become a commodity and models are becoming smaller, open and more efficient. Bonus: ELI5 what part of the above wasn’t inevitable.
-
A Critic Argues xAI Is Shifting Toward Cloud Services
By
–
The other thing that is absurd about $SPCX’s valuation is that Musk just isn’t that good at AI. This new BI article on the chaos within xAI is a case in point. And it’s why he is transitioning to being a cloud provider rather than a frontier leader.