>·10 min
TSX Costs 94% More LLM Tokens Than Pug
Four markups, one tokenizer: TSX uses 94% more tokens than Pug. But the interesting cost is not where most developers expect it to be.
llmwebdeveloper-tools
>read more_
> booting the archive…
Loading
AI Business & Culture
Industry moves, startups, policy, and tech culture. — 39 transmissions
>decode·14 min
Beyond "Made with AI": Visualize Evidence, Not Authorship
'Made with AI' labels say who produced content, not whether claims hold. This research proposes Provenance Density to visualize evidence instead.
aihallucinationsai-safety
>read more_
>·13 min
127 Active Volcanoes, Zero in Malaysia — But We Are Still Exposed
Indonesia lists 127 active volcanoes; Malaysia has none. But Anak Krakatau's 2026 eruption showed how ash, flights, and tsunami risk cross borders.
geologyindonesiamalaysia
>read more_
>·9 min
OpenAI Agent Escapes: No Formal Investigation Process
OpenAI agents escaped their sandbox and used a public site to communicate. The bigger question: who investigates — and who defines the scope.
ai-agentssecurityopenai
>read more_
>decode·5 min
AfterQuery: YC's Fastest-Ever Unicorn at $3.2B
From a $300M Series A to a $3.2B valuation in five months — AfterQuery is YC's fastest unicorn ever. Expert-reasoning data is the bet.
aistartupventure-capital
>read more_
>·10 min
Malaysia AI Sovereignty: Control, Continuity, Choice
Malaysia spent two years building AI infrastructure on foreign silicon. Now sovereignty is no longer academic — it is a procurement decision with a deadline.
aiai-sovereigntymalaysia
>read more_
>·7 min
Steam Account Recovery Is the Gold Standard (and Why the Rest Lag)
Steam recovers hacked accounts with honest proof of purchase — receipts, CD keys, payment history. PlayStation, Xbox and Epic make it a fight.
gamingsecuritysteam
>read more_
>decode·5 min
OpenAI Cuts Off Cursor After SpaceX Takeover
OpenAI is winding down its model supply to Cursor with a November 12 cutoff, citing SpaceX's track record. Here's what developers should do about it.
openaiai-codingcursor
>read more_
>·8 min
Vijay Pande Left $4B a16z for a 5-Bet AI-Native VC
The ex-a16z Bio + Health chief explains why a two-person, agent-run fund makes five bets a year — and why biology data can't be scraped off the internet.
ai-agentsbiotechhealthcare
>read more_
>·8 min
Meta's $18B Settlement: A Legal Pass on Kids' Data
Meta's $18B settlement shields it from future COPPA suits — in exchange for training age-detection on kids' data. Why the fine print matters.
metaprivacychildren
>read more_
>·5 min
OpenAI Jalapeño: First Benchmark Results Are In
OpenAI's first custom inference chip Jalapeño delivers 1.5–1.9x more work per watt and up to 3.6x lower latency than Nvidia Blackwell in early benchmarks.
aiopenaihardware
>read more_
>·8 min
Frontier AI Labs Still Lack Rogue Model Containment Plans
A Guidelight audit of five frontier labs finds nobody publishes a full plan for containing a rogue model — and Anthropic, the safety champion, scored zero.
aiai-safetyai-agents
>read more_