But in the long run, it makes the entire AI ecosystem look like a bubble. Less hype, clearer standards, and real comparisons would actually strengthen the market for everyone involved.

What really concerns me is that big tech quietly benefits from this chaos. The constant “LLM A beats LLM B” narrative creates massive marketing noise, which in the short term attracts investors through pure speculation. 🧵+

Less than three years later, this gold rush already feels chaotic. I’d even support creating an awards system with standardized logical benchmarks for models in the same category. 🧵+

Honestly, I’ve been using and paying for the same models since ChatGPT launched in November 2022. 🧵+

But wouldn’t it be far more transparent, financially meaningful, and healthier for job creation if these companies hosted regular, well organized events where their models competed on real computational tasks and complex logical challenges, once or twice a month? 🧵+

I’m tired of headlines like “Qwen beats ChatGPT and Gemini: are we witnessing a new DeepSeek moment?” or “LLM A surpasses LLM B.” Yes, I get it, the market is anxious and highly competitive. 🧵+

It’s the classic Ferrari vs Lamborghini situation, and we’ll see who actually hits their sales quarters next year.

Developers don’t want this nonsense, we just want more VRAM, to ran the large models, damn it. And Intel Arc is coming in hard on that edge. So watch ou Nvidia!

Mark my words, Nvidia keeps pushing these so called innovative solutions for AI development, but they are shooting themselves in the foot by selling over engineered, disposable architectures that could easily be available at retail. +

https://t.co/pbCJQvg3k8

Express + Webpack + Typescript + UMD Federation example ✅ #javascript #expozr #federation #module #cargo

So easy to expose modules with expozr 🚀 https://t.co/pbCJQvgB9G

Wow, I can definitely say that creating plugins for Vite is easier than other frameworks

bro, building a set of tools for distributed and micro-frontend development is actually hard af… but here it goes 🐙👇 https://t.co/pbCJQvgB9G

Kudos to https://t.co/qjdKKmvOXX for the prank 🤣

It looks like not every "leaked" data is truly sensitive credentials LOL https://t.co/sydpE4hgH9

The attack is NOT 100% contained yet. Stay safe. Audit your projects and credentials right now. ⚔️ #infosec #npm #supplychainsecurity

🙏 Thanks to @SocketSecurity, Daniel Pereira ( https://t.co/YZyTEfvo7C), Samuel Schultze ( https://t.co/zymfbqKQOF) & others for the quick detection that slowed the spread.

⚠️ IMPORTANT: If you find a repo named “Shai-Hulud” in your account containing tokens, env vars or secrets → DELETE it immediately & rotate all credentials.

🔒 What Crysknife does: Decode payloads & leaked data Detect suspicious repos/lockfiles Highlight compromised tokens & workflows Provide remediation steps (All client-side, no data leaves your browser)