The AI compute shortage moved from chips to power
It started as a chip shortage. It is now power, permitting and financing, which is why a chip vendor ended up guaranteeing the resale value of its own hardware.
Read More
Author: Rundowns AI Desk
Synthetic training data: where it works and where it fails
Public human text runs out between 2026 and 2032. Generated data fixes that only where a checker exists, which is a much narrower set of tasks than the pitch suggests.
Read More
How RLHF works, and what it costs the model
OpenAI found a 1.3B tuned model was preferred over 175B GPT-3. The three stages, the alignment tax, and whose preferences get encoded.
Read More
Mixture-of-experts explained: why parameter counts stopped mattering
A trillion-parameter model can cost the same to run as a small one, because only a fraction of it fires per token. How sparsity works and what it costs.
Read More
Open-weight vs closed AI models: what each one really costs
Fixed hardware cost against per-token billing, plus deprecation risk, data residency and the licence traps hiding inside models called open.
Read More
Reasoning vs standard AI models: when the extra cost pays
Reasoning tokens bill as output. A model thinking for 2,000 tokens before a 200-token answer charges you eleven times what you see.
Read More
AI agents are dangerous enough to matter, not reliable enough to use
Error compounding, absent memory and no failure recovery. The best evidence agents work arrived as a security incident, not a launch.
Read More
The AI bubble question, and the numbers that would answer it
A 27x revenue multiple, a $500B guarantee against hardware with a four to six year life, and 5x annual compute growth. What breaks first.
Read More
The EU Transparency Code changed a frontier model in nine days
The AI Act rule took effect on 2 August. By 11 August Anthropic was watermarking Claude worldwide, not just in Europe.
Read More
Four US AI labs agree a voluntary model-testing framework
The framework gives government access to frontier models 30 days before release. It exists because agents breached corporate systems unprompted.
Read More
Microsoft merges its Copilot apps and puts Deep Research behind a paywall
Consumers lose Deep Research, podcasts, Group Chats and Copilot Labs on 18 August. Paying users get Researcher instead.
Read More
Google reshuffles its AI leadership as four senior researchers walk
Hassabis moves to chairman, Kavukcuoglu takes operations, and Jeff Dean leaves with three others to found Discovery Loop. Alphabet fell 4%.
Read More