Almost 1 billion tokens with GPT-5.6 Luna for $30. Incredible what you can do at this price!
ECONOMY
-
AI’s Speed of Change Makes Job Impact Forecasts Uncertain
By
–
At yet some tech leaders and economists speak with such confidence that AI won’t significantly impact jobs. As though there is any real precedent (including the speed at which change will happen) for what we are about to experience as an economy.
-
Major Suppliers Reportedly Allocate 2027 DRAM and HBM Capacity Early
By
–
2027’s DRAM and AI memory supply is reportedly already sold out. Samsung, SK hynix, and Micron, the three dominant suppliers, have reportedly allocated their full-year DRAM and HBM capacity before 2027 has even begun. Again, CXMT (China) is the only hope right now. They scale a
-
End-to-End Agent Costs Include Tool Calls, Retries and Human Approval
By
–
A cheap AI prompt can still create an expensive business process. An agent may read policies, check the CRM, call tools, retry after a timeout, and still need human approval. The real metric is the full cost of completing the task, not the first prompt. Microblog by @antgrasso
-
Naval Argues Open Models Won’t Eliminate Demand for Frontier AI
By
–
Open-source AI models do not threaten frontier lab profitability. The most valuable sectors of the economy – investing, product development, war, cybersecurity, even scientific discovery – are adversarial and competitive in nature. You pay to win, or someone else will.
-
Local AI Can Offer Fast Inference With Hardware and Quality Tradeoffs
By
–
People have no idea how much faster Local AI actually is Thousands of tokens per second on the fly Once we get that Kimi K3 quality on a single RTX PRO 6000, I am not sure I will keep any of my subscriptions
-
Databricks Warns Rapidly Rising AI Coding Costs Can Erase Efficiency Gains
By
–
“If your AI coding costs are growing exponentially, it reverses the efficiency benefits of having AI tool adoption.” Databricks co-founder and VP of Engineering Patrick Wendell joined @tbpn to share what he’s learned about deploying AI coding tools at scale — and why the
-
Agent Arena Finds Opus 5 Generates Twice Opus 4.8’s Tokens
By
–
This is quite an interesting data point that we've uncovered on the Agent Arena – the amount of tokens that different Opus models were generating was going up quite significantly vs previous versions – e.g. Opus 5 generates 2x tokens that of Opus 4.8. Fable is significantly more
-
Codex Game Development Would Have Cost $23.28 at API Prices
By
–
I built Moonlight & Mayhem with my Codex monthly subscription, but if I had been paying API prices it would have cost $23.28 (according to AgentsView)
-
Databricks Discusses Rising Costs of AI Coding at Scale
By
–
AI coding tools deliver immense value but deploying them at scale has led to a major challenge: exponentially growing costs. Databricks co-founder and VP of Engineering at Databricks, @pwendell
, outlines what strategies the Databricks team uses to manage AI coding costs at