Ant Stanley

@iamstan.dev

Tell us a bit about yourself

Good news: some of the US international aid funding has been restored! I graphed it here to put the numbers in context of the scale of the cuts. The article says PEPFAR restoration is still to come. 🤞

Recipient: Gavi
Restored (2026): $600M
Pre-cut US baseline: $300M/yr (FY24 enacted; FY25 + FY26 each appropriated at
$300M, both blocked by Kennedy)
Restored as % of baseline: ~100% of the two withheld years
────────────────────────────────────────
Recipient: WFP
Restored (2026): $800M
Pre-cut US baseline: $4.45B (CY2024, top donor; all donors totaled $9.76B)
Restored as % of baseline: 18%
────────────────────────────────────────
Recipient: UNICEF
Restored (2026): $218M
Pre-cut US baseline: $1.118B (CY2024, largest donor)
Restored as % of baseline: 19%
────────────────────────────────────────
Recipient: Global Fund
Restored (2026): $661M
Pre-cut US baseline: $1.65B/yr (FY24 approp.; peaked $2.0B FY23)
Restored as % of baseline: 40% of one year — and explicitly arrears against
prior-year appropriations
────────────────────────────────────────
Recipient: ICRC + IFRC
Restored (2026): $203M
Pre-cut US baseline: US gave CHF 544M to ICRC alone in 2021 as its top donor
Restored as % of baseline: ~⅓ or less
Josh Busby@busbyj2.bsky.social · 3d ago

Good to see that money for GAVI, Global Fund, the World Food Programme and other health program is now starting to finally flow with expectations for more PEPFAR money as well. Gift link. www.nytimes.com/2026/08/02/h...

Want to get AI to be any good at front end design? Build a first class design system, codify it with Storybook, and force it to use that. No components or design elements on the fly. Let AI assemble, not design.

To every trans Londoner: this city is yours too. Today we celebrate London Trans+ Pride and reaffirm a simple truth: you should be free to live as yourself, without fear or prejudice. You are loved. You are wanted. You belong here.

I built this complex system of skills that build out complex plans that review every piece of code 3 times at a minimum with differing gates. I found with Fable I just need to hand it a spec, and it to plan, implement, then review when its done. A few hours later I come back, its done and works.

conputer dipshit@davidcrespo.bsky.social · 2w ago

“you should do a breakthrough” is also a great illustration that when LLMs actually get good at something, it doesn’t mean they *can* do it with careful prompting, it means they can just do it. we’ve seen task after task in software engineering go from needing careful prompting to “ok now do it”

GPT 5.6 Sol without guardrails autonomously hacking Hugging Face and benchmarks showing Kimi K3 matches or passes Fable on the same day. ... 🫡 good luck to all of us