New VRS invite simulation is now available on HLTV 📊 Stay up to date with the race to PGL Major Singapore 2026 on the dedicated page attached below

Pra quem joga no PC e sente que o Windows pesa um pouco enquanto joga, vale dar uma olhada nisso tem algumas configs aqui que podem ajudar a melhorar seu desempenho 👇

Packed week for Rust. Rustler Weekly is out. 🦀 📊 CSV parsing at 180 GB/s : a VLDB 2026 paper from TUM Munich. 22x faster than DuckDB, 256x faster than single-threaded Rust CSV. Speculative parallel parsing, end-to-end SIMD, zero inter-thread communication. 🔮 fearless_simd v1.0 incoming : SIMD in Rust without writing unsafe. Hardware-width vectors, multiversioning built in, three-year security commitment planned. 🔬 State of SIMD in Rust : the most balanced survey available. Five libraries mapped out. The most actionable detail: put #[multiversion] on loops, use #[inline(always)] for tiny functions. 🚂 Topcoat 0.9 : Rails-style web apps in Rust from the creator of Tokio. Server components, WebSocket push, signal-based reactivity. No API layer, no client state management. 💡 Enum variants as their own types : a short post that will live rent-free in your head. Prove at compile time what kind of value you have. No runtime check, no comment. 🔑 hashify : perfect-hash lookups as a proc macro. 4x faster than phf for tiny maps, 40% faster for large maps. Zero heap allocation. 💼 New Rust jobs added daily : rustler.in/jobs 🗓️ Events : Manchester, Berlin, Munich and more. October kicking off strong. 👉 rustler.in/newsletter #RustLang #BuildingInPublic #OpenSource #TechJobs

Introducing Claude Sonnet 5.5, the second model in the Claude 5.5 family. It’s a clear upgrade over Sonnet 5, runs more than 30% faster, and costs up to 30% less for most work.

While @Cloudflare, I managed the @postgres platform team which supported 175+ product teams. They are known for their shipping velocity. The biggest bottleneck was schema migrations. Just shipped safenotsafe.dev/ to check your Postgres migration is safe or not safe?

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

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".

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-1…

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?

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: piped.video/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/…

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

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/mic…

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/Teachi…

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

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