Petey struggles (@peteystruggles) header
Petey struggles (@peteystruggles) profile picture

ai coding agents, limits. like what you see? follow. don't like what you see? follow.

Berlin, Germany
Joined February 2012
Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)Photo by Petey struggles (@peteystruggles)
I use codex hard enough to break through $200 plan and I got tired of clicking though menus so I built Codex Limits - it tracks burn rate, reset windows, banked resets, tells you when your "infinite" plan may actually run out i'm building in public for the rest of August, follow for honest thoughts, practical fixes, small tools that make working with agents better
1
1
10
1,977
willing to sell a kidney for 2x m5 ultra 512gb
Today, we release the weights of a new and improved Le Chonk, advancing the open frontier for agentic coding and cybersecurity. 💻 Built to build: Le Chonk ranks #1 on @cognition’s FrontierCode among open models, and sets a new benchmark for Pareto efficiency on @artificialanalysis. 🛡 Built to defend: State-of-the-art in cyberdefense, put through the paces by the world's leading cyber vendors, including @CrowdStrike, @cloudflare, @rubrik, @zscaler, and others. Built for sovereignty: Make it yours. Download the weights on @huggingface , customize it with Mistral Forge, and deploy it wherever you need it. Read more from our team and customers 👇:
22
two weeks from argon to carbon. google is shipping like a startup again. and this is great news for everyone
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model. Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards. As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
1
1
33
Petey struggles retweeted
Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago. Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned.
1,236
1,155
862
15,862
1,312,913
a perfect sidekick for literally everyone
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵
6
this + infra + android
Google is down bad for no apparent reason So today I have been loading up on $GOOG stock (well $GOOGL but they're mostly the same) for in my eyes very cheap I just Facetime'd with my dad and we talked about ChatGPT and he answered "yes I cancelled my ChatGPT because it's on Google now" I asked what the jackfriks he meant He said "so Google Search has AI now and I trust it more because it links back to sources etc" My survival thesis is we have to be co-owner of BigTech and BigAI while they become to sole owners of exponential intelligence which will replace a lot of our businesses and which they will then sell back to us I'm doing what I'm saying
11
kimi, welcome back
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
13
man, Claude just gets me
12
I love the work you did with aligning the buttons, dear @BetterStackHQ 10/10
27
welcome to the new normal not that long ago frontier labs wanted to regulate ai in a way that would have locked out open-source models. and they wanted to monitor themselves. not everyone can access cyber-sec models that labs offer. but because open models are still here, anyone can defend themselves against bad actors
they’re calling it the black October lmao. > node.js: security vulnerability. > PHP: critical RCE vulnerabilities. > tensorlake: npm package infected. > liteLLM: malicious PyPI + Docker releases > Railway: customer secrets exposed > zeabur - infrastructure breached. > OpenCode: AI API keys exposed. > OVHcloud CLI: official binaries backdoored. > RankMath PRO: malware inside official plugin. > TeamUpdraft: customer data compromised. > ConnectWise: cloud backup identities compromised. > SubQuery: backdoored npm package. > GhostAction: GitHub repos with malicious workflows (related to the broader Shai-Hulud/ChainDrop activity). > eMudhra: certificate authority keys compromised (with downstream impact on Tailscale). this some complete madness bro. 1/3 of the internet is fucked.
23
I like how every lab wants to offer their grok bot which is just a simplified and lobotomised hermes not that it's a bad thing - I truly enjoy grok bot. I would probably enjoy dot too. I am quite certain that duet is going to be a blast as well it's soooo good to have options
We’re opening a research preview of Claude Duet, a new app where Claude works as a proactive agent. It comes to you with what needs your attention, completes tasks you ask of it, keeps working when you step away, and checks with you before it acts on your behalf.
15
Petey struggles retweeted
50
364
33
3,066
76,386
one step ahead by not using PHP
PHP fixes out. LETS RACE
402
seriously impressive
Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵
24
okay, this explains a lot thanks, OpenAI
what happened to codex's usage today? first I noticed the sudden drop in remaining usage (97% to 35%) and now every minute of 6.1 sol is 1% of weekly usage. like, what? how?
11
I'm amazed how easy it was to setup remote control for all my projects on a vps with Claude app I can use literally any device and pick up work immediately this is currently the best app to do work
1
2
24
gpt $20 vs claude $20 gpt-6.1 medium efficiency is interesting. just a simple task checking out email and if containers are live - 2% weekly burned, 13% 5-hour limit up the furnace in the meantime I literally have 24/7 opus 5.5 medium and my allowance is still there
2
2
40
hey chat, what are you shipping this weekend?
1
18