erik.eth 🛡

It’s been incredible to see x402 go viral this week! For anyone new, welcome to the movement. I designed x402 to be the HTTP of money: a universally accepted, neutral, core primitive to help make the internet ready for the AI era. I can’t wait to see what people build 🛠️

Of all the "manage many agents" tools i've tried @orca_build is definitely the best. @zeddotdev for editing code / pair programming with an agent for high precision things, Orca for swarming. github.com/stablyai/orca

@nsharma0813 It’s dated (website is open source, would love if some submitted a PR to make it live)

hey @programmer would you know if this data on x402.org accurate/realtime or dated?

@artman What is currently better? It does feel like a local maxima but also is clearly better than 1 PR

Stacked PR’s are an outdated, inefficient solution to a problem that only exist because the root problem hasn’t been addressed in GitHub: review tools need to do a better job at help the reviewer work through large code changes. It’s a local maxima.

Stacked diffs are now more useful than ever. Far easier to review LLM code if it’s broken into stacked diffs.

@sameenkarim @bilaltwovec @github Great launch congrats! Web app performance seems better in stacks than it did before!

The launch of Stacked PRs is one of the biggest releases in @github history, touching almost every service we run I wanted to share some details about how we built it and everything that went into it Strap in folks, this is gonna be a long one 🧵

@gaurangtorvekar @ChatGPT @linear It seems to be ChatGPT tasks ui being unauthed

@programmer @ChatGPT @linear Oh damn! What if you ask Codex to pick up tasks FROM Linear and finish them one by one?

fyi all @ChatGPT Codex tasks you start from @linear are default public & leak your internal repo + branch names. Doesn't seem to be a way to force them private.

Awesome write up from @fastly today - they are bringing x402 payments to the edge ⚡ fastly.com/blog/automated… This new demo shows how Fastly Compute can enforce payment before a request ever reaches the origin, letting agents pay for APIs and digital resources automatically with @USDC on @Base. A powerful pattern for metered APIs, premium data, crawlers, and machine-to-machine services.

top 10 must try agent services for this week, part 2 ​ (all pay-per-use instead of subscriptions) -api.arkm.com (by @arkham) gives agents pay-per-request Arkham wallet intelligence, token flows, and market data via x402! start using: x.com/agentcashdev/s… -agentres.dev (by @blauyourmind) lets your agent book Resy reservations and only pay when the booking succeeds. demo here: x.com/blauyourmind/s… -api.purch.xyz (by @trypurch) lets agents search, shop, and buy products or vault items over x402. demo here: x.com/DavidPawlan/st… -api.locus.report (by @velinus_sage) gives agents cited property/context reports from public records, policy, permits, and risk data. check out the type of analysis that can be done @ api.locus.report/api/reports/lo… -laso.finance (by @LasoFinance) lets agents buy prepaid debit cards with USDC via x402. used it this weekend so my agent could order me starbucks lol. check it out here: x402scan.com/server/2be7b63… -stabledetect.dev (by @OnchainLu) gives your agent AI/deepfake detection for text, images, and video via pay-per-call endpoints. demo here: x.com/OnchainLu/stat… -modelprices.xyz (creator unknown) gives agents normalized LLM prices, limits, and price-change feeds for 2,000+ models. demo here: tryponcho.com/s/b31853f2-5d0… -signals.neilkpatel.com (creator unknown) offers high quality, per-call equity research signals. see how i used it: tryponcho.com/s/8a2a9793-eb5… -cpi-report-us.underscoredone.com (creator unknown) gives agents official US CPI and core inflation reports from BLS w/ plain-English summaries. demo here: tryponcho.com/s/433fbaf1-4dc… -know.fast (creator unknown) turns any URL into a durable cited knowledge page and agent-ready markdown packet. demo here: tryponcho.com/s/c9c42eb1-91a…

gotta say, grok build with 4.5 feels insanely fast

@niallohiggins I’m assume token efficiency will also get solved for open models eventually, but even at 3x the tokens glm 5.2 is ~1/3 the cost. I don’t think it’s necessarily about this generation of models, a year from now an 120b model will be roughly as good 5.2, and it’ll cost even less.

@programmer You also have to look at token efficiency. If GLM 5.2 has to output three as many tokens as GPT 5.5 for the same agentic task, does it matter if it costs half the price?

Just wait…

biggest question of our time

@programmer I raising money for a 25 disc CD changer

every github repo being backed by a CD would actually explain a lot about github's performance & uptime

This is both cool and also wtf. the people want better site performance and uptime

Charge for any asset protected by Cloudflare. In stablecoins. Over the x402 protocol.

It feels like the era of minimalism is ending. return to maximalism.

Agents disconnect you from your users. Ads won't work the same in the age of ai. x402 solves this.