stridell

@_trish_xD english

first time you wrote hello world - what language did you use?

Giveaway time! Win a Claude Max 20x subscription for one month. To enter > follow @BleapApp > RT and like this post Winner will be selected in 48 hours. (fyi, you get 20% cashback on your claude, chatgpt and gemini subscriptions when using a Bleap card) Download the app via the link in our bio today and activate your virtual card in minutes.

@KeyTryer Benchmaxxed

How is Gemini 3.1 simultaneously the best and worst model?

@EtherCoins @QuixiAI @__tinygrad__ No shit there’s no CUDA that’s proprietary nvidia stuff

@QuixiAI @__tinygrad__ The board is useless. No CUDA, no competitive bandwidth, no adoption. You're welcome.

Intel B70 finally makes a truly competitive move. 32 GB vram for < $1000 No matter how bad the software stack is, the sheer vram / dollar ratio will drive the community to fill in the gaps @__tinygrad__ Intel tinybox?

@kimmonismus Yeah, no. It absolutely sucks. Tried it but literally couldn’t even finish basic tasks.

Nemotron remains completely underrepresented. NVIDIA has cooked up a storm, and most people are still ignoring it.

@de3dsoul subnautica

What game is this?

@callmecelebi @TeksEdge @intel CUDA is proprietary Nvidia tech, they can’t if they want to

@TeksEdge @intel It impacts my heart 😢 Joking aside, hope they will consider CUDA for next line or model.

🚨 Exciting Local Inferencing News! @intel just dropped the Arc Pro B70 a serious new opportunity for local AI inference! 🔥💥 💰 Price: $949 📆 Available: Starting today (March 25, 2026) Key Specs: 32GB GDDR6 VRAM 📦 (608 GB/s bandwidth ⚡) 32 Xe2 cores + 256 XMX engines 🧠 Up to 367 peak TOPS 🚀 TDP: 160–290W (Intel version ~230W) ⚡ Intel claims massive gains 📏 2.2x larger context windows ⚡ 85% higher token throughput 🏎️ 6.2x faster Time-to-First-Token 💵 Better token-per-dollar vs RTX Pro 4000 Equivalent to? 🧩 Strong competitor to the NVIDIA RTX 5000 Ada / Blackwell (32GB class) for local LLM inference — especially on Linux setups Excellent value under $1k for AI/agent workloads 👀 Worth it over used 3090s? Or still sticking with NVIDIA? Drop your thoughts below 👇

@Trae_ai @OpenAI about time?

GPT-5.4 is now available in TRAE. With upgraded reasoning abilities and enhanced image comprehension, @OpenAI GPT-5.4 delivers more precise and context-aware outputs. Elevate your development with improved efficiency and deeper web search capabilities in TRAE now.

@zeddotdev Did you also generate the marketing with it? We didn’t just x. We y.

Zeta2 is here. 30% better acceptance rate than Zeta1. 200x more training data, LSP-powered context, faster predictions, open weights. Try it now in Zed. We didn't just improve the model. We rebuilt the entire data pipeline behind it: zed.dev/blog/zeta2

@testerlabor Grok 4 has 3 trillion params? why is it so stupid then

Grok 5 is training on Colossus 2, the world’s largest Supercluster and is expected to have 6 trillion parameters - roughly double that of Grok 4. The most exciting and most powerful outcome is most likely.

@RevealingSkunk @SOSOHAJALAB 27B params vs well over 1T btw. The density is the cool part.

@SOSOHAJALAB Nah it's unbelievably stupid, I asked it to replace a refactored method call in tests, it simply substituted the name without inferring parameters and updating mocks. Real claude did all of that easily

Guys.. this model is just crazy. If you have just less than 48gb vram, just try the 8q gguf format. Feels just like opus! Tool calling is working smoothly!! Appreciate for this! (Hf and qwen!!) huggingface.co/Jackrong/Qwen3…

LiteLLM HAS BEEN COMPROMISED, DO NOT UPDATE. We just discovered that LiteLLM pypi release 1.82.8. It has been compromised, it contains litellm_init.pth with base64 encoded instructions to send all the credentials it can find to remote server + self-replicate. link below

Great

@DavidOndrej1 it’s a bug, unless you’re trying to dictate a message >10 min long

i’m paying $200 a month for ChatGPT pro yet I have 10 minutes of dictation only?! what the fuck

do people really believe this lol

@Yampeleg Last gen sonnet no thinking? Why

Current stack: - Pi for everything. - GPT-5.4 for everything code. - Gemini-3.1 for design/brainstorming. - Sonnet 4.5-no-thinking for openclaw. - GLM-5 for parallel swarms. - Opus 4.6 for everything else. Currently Testing: Minimax-2.7.

@1kartikkabadi1 @0xSero Why are you paying for it still?

@0xSero Shit. I might just have to cancel yt premium finally

First they removed dislikes from YouTube, now they removed Likes. No more public signal on what matters. Making another braindead prediction, in 12 months we won't see views anymore. In 24 months we won't subscriber counts anymore.

@Amank1412 Seen so many people disliking the chart. What’s wrong with it?

>be cursor >fork vscode >cross $1B ARR faster than ever >openai tried to replace you with codex (and bought astral just in case) >drop your own model anyway >make your own benchmark >rank yourself no.1 on it >make a criminally bad chart (and we still don't know what cursorbench is)

@BentoBoiNFT @GG_Observatory Nah he just types in newline, you gotta render the post in your head to understand

@GG_Observatory Is this ai

Why would anyone choose OpenClaw vs Claude Code? Claude now has: • Discord/Telegram integration • Cron Jobs (/loop) • 1M token memory • Webhooks to phone • Can run 24/7 on any Computer or Mac Mini This covers 95% of what people actually use OpenClaw for with better security and easier setup The only reason to stick with OpenClaw is if you want a multi-agent setup. That's the only difference I could think of Going to stick with OpenClaw for now because of this, but the gap is almost at zero