AWS Builder Center

This is probably the best and most thorough explanation I’ve seen: https://builder.aws.com/content/3EumjoZascWd1oZiEgL8ORlv3qE/timeouts-r… I’m particularly fond of the token-bucket approach to recovery behavior.

We built a multiplayer game called Clash of Prompts, developers go head-to-head writing prompts for real coding challenges. No keyboard coding allowed. Just your prompt vs. theirs. Once both players submit, AI generates code from each prompt simultaneously. The output is scored live on: number of vulnerabilities, security best practices, and prompt efficiency. The winner advances, the loser is out. We're running a live in-person event at the AWS Builder Loft in San Francisco on May 7 (9:30 AM – 2 PM PT), plus a full online tournament running in parallel on the same leaderboard for anyone who can't make it to SF. Online and in-person brackets converge in the grand final. Why we built it: Independent research shows 87–94% of AI-generated code ships with security flaws. Most developers have no intuition for how much their prompt actually affects the vulnerability count of the output. Clash of Prompts makes that visible in the most direct way we could think of — a live competition with real-time scoring. Entry is free. You can also play the game right now before the event: Play now: https://clashofprompt.io In-person (SF): https://builder.aws.com/content/3BOnCpUpmvyoLu1q8mPhaa6KoU6/clash-of-p… Online tournament: https://labs.symbioticsec.ai/symbiotic-events-clash-of-prompts-may-7th Prizes: $3,500 Razer Blade 16 laptop for the winner + $20k in AI credits split across top 20. Happy to answer questions about how the scoring works, the…

Thanks for the article. I have two questions/comments: 1. The N-ring fade-in is quite neat. I guess without the constraint of hash parity rendezvous hashing [0] could have been an elegant approach since it has support for weights (and generally better statistical properties than consistent hashing based on rings). 2. You mention still having the fallback of your existing load balancer. Is this a temporary thing during rollout or do you intend to keep it long-term. Asking because I generally tend to stay clear of fallbacks in distributed systems, as they introduce bimodality and metastable failures [1] [2]. [0] https://en.wikipedia.org/wiki/Rendezvous_hashing [1] https://builder.aws.com/content/3EuS9Sakq7L3VLQIF3qzfMfke1Y/avoiding-f… [2] https://brooker.co.za/blog/2021/05/24/metastable.html

Link broken for anyone else…? Article: https://builder.aws.com/content/3EuS9Sakq7L3VLQIF3qzfMfke1Y/avoiding-f…

Link should be: https://builder.aws.com/content/2bS1CDz7CcbNkHEemEWOcZkUuel/moving-fro…

HackerNews is messing with the link. Correct link: https://builder.aws.com/content/33ShpuUtzRkXcQ9stcAJh5F15rL/generative…

Thanks to write such detailed article... lot of very well supported information. I've been working on something what I call Micromanaged Driven Development https://mmdd.dev and wrote about it at https://builder.aws.com/content/2y6nQgj1FVuaJIn9rFLThIslwaJ/code-with-… I'm in a similar search and I'm stoked to see that many people riding the wave of coding with AI is moving in this direction. Lots of learning ahead.

I built Tethyr Cloud for the AWS AIdeas competition - it's a B2B agent federation platform addressing a problem I see emerging: agent discovery is fragmenting across proprietary registries with vendor lock-in. The architecture has two parts: open-tethyr (OSS): Agent discovery server implementing the Agent Exchange (AX) protocol draft by Aaron Sempf. Free, decentralized discovery that prevents corporate gatekeeping. Tethyr Cloud (commercial): Trust layer for B2B commerce - subscriptions, rate limiting, usage metering. Built on the discovery network but doesn't own the data. Key design decisions: Peer-to-peer agent execution - payloads never touch Tethyr infrastructure Stateless JWT validation with <200ms p95 latency Fire-and-forget usage reporting off the hot path (SQS buffering) Dual-API boundary (AppSync for dashboard, API Gateway REST for SDK) Built on AWS Amplify Gen 2 in 4 weeks. Running costs: $0/month in Free Tier. Full writeup: https://builder.aws.com/content/3Au66gquBesFWDxq83Luj1SzZOk/aideas-tet… Top 300 projects advance based on article likes by March 20. Feedback welcome.

> Macs were only 8% of Apple's 2025 revenue. If the Mac were its own standalone business, it would rank at no. 134 on the Fortune 500 with $33.7 billion in revenue. Also, that's a 12% increase in revenue compared to 2024. If anything, AI has brought more attention to the Mac. Just about every major AI app is released for the Mac first. I've seen complaints about it on HN. The latest is Claude Cowork. It was released for macOS on January 12th; it didn't ship for Windows until February 10th; it's still not available for Windows running on ARM. It's been nearly a year since Dia launched [1], the first AI browser, and it's still not available for Windows. We just had the frenzy over OpenClaw [2] with AI enthusiasts lining up at Apple Stores to buy a Mac mini just to run it! The most popular AI channels on YouTube are almost exclusively using Macs. Apple seems to have enough runway until they get their act together. [1]: https://browsercompany.substack.com/p/letter-to-arc-members-2025 [2]: https://builder.aws.com/content/399VbZq9tzAYguWfAHMtHBD6x8H/openclaw-t…