André Batista @klunkySQL

Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.

This is cool try.tldraw.com

Siddhartha @sidposting

a week ago

This is cool try.cloudflare.com

161 151 3K 1.9M 5K

This article is such a treat. 1) A reminder that at scale, even "tiny" optimizations make a huge difference. The few optimizations made here saved them literally millions of $$ of RAM. 2) So many fun little CS and math things. Consistent hashing. Birthday paradox Storing struct fields as a byte array. Stats! Side note: shocker I read an entire article with 0 mentions of "database".

Cloudflare @Cloudflare

a week ago

100 TB of RAM, saved by shrinking a consistent hash ring. The last 90,000 hashes per server were buying 0.7% load balance improvement. Math said stop. We stopped. blog.cloudflare.com/saving-100-tb-…

44 157 2K 448K 1K

100 TB of RAM, saved by shrinking a consistent hash ring. The last 90,000 hashes per server were buying 0.7% load balance improvement. Math said stop. We stopped. blog.cloudflare.com/saving-100-tb-…

Desde que comecei minha maratona de IA em fevereiro deste ano, estou mantendo 43 repositórios (36 open source), mais de 1.6 milhão de linhas de código. ai-memory, ai-usagebar, ai-jail, FrankMD, llm-coding-benchmark, distrobox-gaming, etc Acompanhem aqui: github.com/akitaonrails O que vocês andam produzindo?

keyboard_arrow_left Previous keyboard_arrow_right Next

I don't use Codex anymore. I use Claude Code to plan, Codex to execute only through the Codex MCP managed by Claude, and a dedicated environments so they can keep working when my laptop closes. My notes and skills live in Obsidian, so I can switch tools and devices without rebuilding that context. My setup walkthrough: youtu.be/KeyqBiQYMok

Post novo: "Falando um pouco sobre minhas Skills de IA" Esse é mais curtinho, só dando pitaco sobre as skills que eu uso no meu dia a dia e algumas dicas e recomendações pra quem ainda tem dúvidas. akitaonrails.com/2026/09/17/fal…

Publishing a Kubernetes SIG's Images to registry.k8s.io - kubernetes.dev/blog/2026/09/1… #Kubernetes

Again, read the comments, not the OP. 80% margins means they keep 80 cents on the dollar which sounds great in theory but then you have to add in what revenue they share with their distribution partners and all of their operating expenses and that 80% goes down real quick. You cannot overlook this because they absolutely have to spend that money. And by law they have to pay their employees first before all other vendors and operating expenses. COGS = cost of goods sold Also see the line about stock based compensation which goes back to what I was saying yesterday about the AI researchers and why they're jumping ship now. I said that they would only do so if they were afraid they were going to be out of a job or if the compensation they were promised (i.e stock) is not going to be worth what they were initially told it would be. Depending on their contracts they may be able to exercise pre-IPO shares. This now seems to be lining up with my suspicions. You don't leave a project you signed up for in anticipation of stock-based compensation ahead of actually receiving that compensation unless something is very wrong with the compensation. It has absolutely nothing to do with the product being alive because it isn't. It has to do with these AI researchers wanting to continue to pay their bills and eat each day. Even if they can't pre exercise, they are absolutely being paid to leave. No one is doing this for free and there's no such thing as a free lunch. All those days in the boardroom and I actually learned something! 😉 Anyway this is going to be a glorious dumpster fire and I can't wait. If you thought people were pissed over the FTX thing wait till you see this. This is going to be a spectacular rug pull. Especially when they start getting into how much all of this hardware costs and that it has to be replaced every two to three years versus how many paying customers they have and the expected revenue, which doesn't come anywhere close to the hardware costs. That number has continued to decline not increase, by the way. In my response here I'm going to post the poly market thread because that one's hilarious, too. Fintech is responding. These are the people that understand financial statements (So do I, it was part of my job as head of finance 😉) and can tell you what this looks like before they IPO. Pay attention to what the financial experts are saying. What they are saying is what myself and everybody else here as cybersecurity experts are saying, which is that this is going to be a disaster. Also Anthropic is trying to get ahead of the bad news in their financial statements which are about to become public by frontrunning these warnings couched in clever accounting trick phrases. 80% margin.😂 For context, Apple and Amazon have around 50% gross margin.

Jukan @jukan05

2 weeks ago

“Anthropic’s gross margins are above 80% before accounting for revenue shared with distribution partners, including Amazon, and the cost of training its models.” Wow.

396 336 5K 5.1M 1K

Kubernetes v1.37: Pod-Level Resource Managers graduated to Beta #devopsish #kubernetes kubernetes.io/blog/2026/09/1…

NEW TODAY: NIST IR 8587 : Protecting Tokens and Assertions from Forgery, Theft, and Misuse: Implementation Recommendations for Agencies and Cloud Service Providers (final) csrc.nist.gov/pubs/ir/8587/f…

BREAKING: Microsoft has just announced a Windows and Surface event for October 7th. The event will take place in San Francisco, with Microsoft CEO Satya Nadella, Windows chief Pavan Davuluri, and Nvidia CEO Jensen Huang. Full details 👇 theverge.com/news/994714/mi…

Blizzard’s inability to solve the login problem on their product launch day reminds me of the classic Forster essay: The Machine Stops cs.ucdavis.edu/~koehl/Teachin…

The lifecycle of a sharded Postgres query — PlanetScale #devopsish planetscale.com/blog/the-lifec…

Here it is: Performance Improvements in .NET 11 devblogs.microsoft.com/dotnet/perform…

Kubernetes v1.37: Memory QoS Graduates to Beta #devopsish #kubernetes kubernetes.io/blog/2026/09/1…

@atoonk cidadaospelaciberseguranca.com/2026/09/15/apa…

Multiple drops in traffic for Altice's MEO (AS3243) in Portugal in the past 24hrs. Traffic drops coincide with customer networks routing away from MEO's backbone network (AS8657). Perhaps an internal routing problem like what occurred for Rogers (2022) or Optus (2023)?

keyboard_arrow_left Previous keyboard_arrow_right Next

I wrote a blog post about exploiting Logi Options+ for SYSTEM shells. blog.amberwolf.com/blog/2026/sept…

Huge week for Rust. Rustler Weekly is out. 🦀 🏆 Microsoft makes Rust a Tier-1 language : same tier as C++, C#, and TypeScript. New compiler backend connects rustc to MSVC. Cross-language inlining with C++, binary hardening, hotpatch support. Running in 100+ internal Microsoft repos today. ⚡ A single Clippy lint was eating 25% of total runtime : it's now 3133x faster. Less than 200 lines of code. The fix was moving the check before macro expansion. 🔬 Seven async runtimes, seven different outputs : a Brown University paper maps nine design dimensions hiding behind two shared keywords. The same twelve-line program prints different things in every runtime. 🐛 Rust debugging survey : 2,300 developers, over half don't use a debugger at all. Print debugging wins. The biggest pain points: enums showing implementation details, not variants. 📟 Which MCUs actually have good Rust support : a practical breakdown by use case. RP2350 is the most underrated right now. 🔧 ESP32 as an SWD programmer for STM32 : the full SWD protocol implemented in no_std Rust. Starting from a Reddit post. Anduril is hiring across multiple Rust roles this week. rustler.in/jobs 🗓️ Events : Seattle, Prague, Austin, Aarhus. Late September. 👉 rustler.in/newsletter #RustLang #BuildingInPublic #OpenSource #TechJobs