Jared Duggan

@OpenAI shutting down SORA...... I honestly didn't see that coming.

@altryne Yes both Monday and Tuesday (today) blew through usage like never before

My feed is showing me a bunch of folks who tapped out their whole usage limits on Mon/Tue. Is this your experience? Please comment, I want to understand how widespread this is

@10xtin @Al_Grigor Also, its about communication. If they are reducing their token usage on Max plans let users know, before it happens!

@Al_Grigor You've got 5000$ of inference for 200$ and are now unhappy with them cutting it to make it more profitable? 🥲

I hit my limits very quick this week - even with 20x pro plan. It makes my claude code unusable A good reason to do more stuff with Codex!

@Al_Grigor This has been happening to me since yesterday also. Either there is a bug in claude code or @anthropic has significantly reduced their token allowance on Max plans and has not announced it.

I wouldn't say he is wrong, we all have different preferences. I have tried Minimax models in Openclaw and was so unimpressed I likely will never use them again. They literally broke configs at least once a day ans never followed memory or rules. Qwen 3.5 27b was better but felt sluggish to me, acceptable but not optimal.

Peter is wrong He needs to try MiniMax M2.7 and Qwen 3.5 27B in OpenClaw before making these comments

@callebtc Two very loaded questions there!

anyone tried openclaw? do you like it?

@chris__sev Did you set up dangerously skip permissions? Otherwise how are you avoiding approving Bash commands and virtually every other command that needs approval?

Claude Code remote control always disconnects Claude Cowork Dispatch asks for permissions on every chat Claude Code Channels is finally the most OpenClaw thing. It's been fantastic. Doesn't disconnect. Responds fast. Never bothers you for permissions. I think we have a winner 🤌

Cool new feature, you are gunning for OoenClaw! One feature request: allow MCP channels (like Telegram) to act as an approval gate for tool calls. Right now there's no way to route bash/command confirmations through an MCP server — would love to approve or deny commands from my phone via Telegram instead of needing terminal access.

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

@MatthewBerman Probably GLM 5

What’s the best open source model for coding right now? Any size 😆

@SawyerMerritt Compare this to model y

BREAKING: Rivian has officially announced pricing and specs for their new smaller R2 SUV. Here are the trims: R2 Performance (with Launch Package): • Starting price: $57,990 • Deliveries start Spring 2026 • Range: 330 miles • 88 kWh battery • 0-60mph: 3.6s • 656 hp • dual-motor • 10% to 80% charge in 29 mins • 9 speaker sound system • Semi-active suspension • Rear drop glass • Matrix LED headlights • Heated/ventilated front seats, heated rear • 186" long (15" shorter than R1S) • 9.6" ground clearance • Towing capacity: 4,400 lbs • Native NACS port • 21" wheels • 90.1 cu-ft total storage • The Launch Package includes: lifetime subscription to Autonomy+ ($2,500 value), tow package, and Launch Green Color available. R2 Premium: • Starting price: $53,990 • Deliveries start late 2026 • Range: 330 miles • 88 kWh battery • 450 hp • dual-motor • 0-60mph: 4.6s • 10% to 80% charge in 29 mins • 9-speaker sound system • Matrix LED headlights • Heated/ventilated front seats, heated rear • Rear drop glass • Fewer drive modes • No semi-active suspension • 186" long • 9.6" ground clearance • 20" wheels • 90.1 cu-ft total storage • Towing capacity: 4,400 lbs • Native NACS port R2 Standard (RWD): • Starting price: $48,490 • Deliveries start in 2027 • Range: 345 miles • 88 kWh battery • single-motor • 350 hp • No all-terrain drive mode • Only all-black interior available • 5-speaker sound system • 0-60mph: 5.9s • Tailgate window glass doesn't drop down • 186" long • 90.1 cu-ft total storage • 9.6" ground clearance • Towing capacity: 4,400 lbs • Native NACS port Additional R2 Standard trim: • Starting price: $45,000 • Deliveries start late 2027 • Range: 275 miles • Other specs: TBD More photos in thread below:

@Hunterjones @DallasAptGP No GPT here. Just someone solving real issues I face on a daily basis.

@DugganJared @DallasAptGP Thanks Jared GPT

My attorney and I were just comparing notes on AI. His comment: “HarveyAI (considered best in class) is what I used at my large firm. $2,500/seat/month. Claude CoWork (with legal plugin)…..$500/month for 5 seats (2 premium). Claude is just as good as Harvey.”

Great point Hunter, and you're right — the models are evolving at an insane pace. But that's actually the best part for platforms like Logos AI. We're model-agnostic. When a better model drops, we plug it in. Our users have access to Claude 4.6, Gemini 3.1, etc the moment they were available. The platform gets smarter automatically. But here's what a raw model subscription will never give you — no matter how advanced it gets: A ChatGPT or Claude subscription doesn't organize your documents by case. It doesn't track your deadlines. It doesn't let you chat across 200 pages of medical records tied to a specific matter with persistent context. It doesn't generate research memos with real case citations formatted for legal work. It doesn't track AI costs per case for client billing. It doesn't version your contracts or run risk assessments. And it's not HIPAA-compliant. Models are the engine. A legal AI platform is the car. You can have the most powerful engine in the world, but without the steering, the navigation, and the framework around it — you're not getting anywhere efficiently. The companies that won't last are the ones built around a single model or selling access to something you can already get with a subscription. The ones that will last are the ones solving real workflow problems on top of those models. That's what we're building.

@DallasAptGP The AI space is changing so fast I can’t imagine how long these companies can last. The models and capabilities of regular subscriptions is more advanced every couple months.

Barrett, this is exactly why I built Logos AI. As a practicing attorney, I got tired of the status quo — $2,500/seat/month platforms like Harvey that lock you into long-term contracts with minimum seat licenses before you even know if it fits your workflow. But here's the thing — Claude CoWork with a legal plugin is a great general AI tool, not a legal platform. It doesn't give you case management, deadline tracking, or organized matter-based workflows. You can't upload a client's medical records, deposition transcripts, and insurance correspondence and then chat across all of those documents with full context tied to a specific case. There's no per-case cost tracking for client billing. No document versioning. No practice-area-specific workflows. It's a chatbot with a legal skin — not a legal operating system. That's what I built Logos AI to be. A platform designed BY a practicing attorney, FOR attorneys: ✔ AI-native case management with task extraction and deadline tracking ✔ Intelligent document chat across your entire case file with full context ✔ 12+ page legal research memos with real case citations ✔ Contract drafting with risk assessment and compliance checks ✔ Multiple AI models — Claude, Gemini, Llama, DeepSeek, Mistral ✔ HIPAA-compliant from the ground up ✔ Per-case AI cost tracking for transparent client billing ✔ Month-to-month. No seat minimums. No long-term contracts. Starting at $179.99/mo. Legal AI shouldn't cost more than your associate — and it shouldn't be a generic chatbot dressed up as one either. trustlogos.ai

@Shpigford @openclaw You need to be careful with cheap small models. They are susceptible to prompt injection and also screw up openclaw settings on the regular. There is not great answer to your question at this time.

what's the best low-cost model for use with @openclaw? want to do some testing around running an instance as inexpensively as possible. trying to fine the right balance between cost and functional.

This is exactly right, and it's why I built Logos AI (trustlogos.ai). As a practicing lawyer who uses AI, I can confirm that tools like Logos make me significantly better at my job, help me work faster, and give me a real edge that lawyers who don't use AI simply won't have. Every lawyer will need AI to keep up. When I demonstrate the power of the platform to other lawyers who aren't using AI, they honestly cannot believe it.

You guys don’t get it yet. Everyone keeps saying AI is going to replace lawyers. I don’t think people understand how this actually plays out. Let’s say you use AI to draft a contract. The contract misses something important. A year later it costs you two million dollars. What do you do? Right now, you sue your lawyer. In the AI world, you’d sue the AI company. Two things can happen. Option 1: The AI company has liability for legal advice. If that’s the case, every AI company will immediately stop letting consumers use AI for real legal work. The liability risk is massive. Option 2: The AI company has no liability because of disclaimers. If that happens, every state bar in the country will say consumers are being exposed to unregulated legal advice and call it the unauthorized practice of law. And they’ll shut it down that way. Either path leads to the same outcome. Consumer AI will be limited to generic “Wikipedia-style” legal information and LegalZoom level document prep. But the real AI tools? Those will live inside law firms. Lawyers will use them to move faster, analyze more data, and run way more matters at once. The M&A lawyer doing 5 deals at a time will do 50. Trial lawyers will run far more cases simultaneously. The idea that AI replaces lawyers probably dies. The more likely outcome is that AI supercharges the best lawyers and makes the profession even more profitable than ever.

@Zeneca So far performs pretty well IMO. I always use GPT in conjunction with Opus. Honestly I dont think anyone can say which is better, they both are great models.

GPT-5.4 is out now, can’t wait to give it a try Anyone used it yet for coding? How does it compare to codex or opus 4.6?

Posts like this are click bait and people should stop. You cannot run Opus level models on Mac Studios yet. So this comparison makes no sense. Now I'm a big proponent of self hosting and running whatever models you can locally, but be honest and truthful for people who dont know.

You can pay $200/m for Claude for eternity Or buy a Mac Studio for $166/m and run your own model for free Best part? You own it after 12 months at 0% interest: > M4 Max > 36GB Memory > 512 GB SSD Hmmm…

@VadimStrizheus As a founder, neither. Openclaw has some utility for preconfigured daily code reviews. Cowork is just a wrapper for claude code so most founders I'm sure are using CC.

As a founder, which one do you prefer? 1. Claude Cowork 2. OpenClaw

@levelsio What model do you have your tool calls set to use, I think it defaults to Haiku if you don't change it yourself

Claude Code with Opus 4.6 was so dumb today I finally had to write my own code again A sad state of affairs 🥹

everybody on your timeline is lying to you about OpenClaw I started using it pretty much the first week it came out, and I was impressed and went deep into the rabbit hole its not the 24/7 AI agent that everyone is making it seem like, not even close if you need to build a website or code something, just use Claude Code or Codex directly. if you need to generate images, go straight to Nano Banana or Higgsfield. Using OpenClaw for these things is like taking a detour through 5 cities to get to the house next door the people telling you "just install OpenClaw and let it run your business" are either lying to you or havent actually used it for more than a demo it breaks constantly. it forgets context. it takes 10x longer than just using the right tool for the job. the experience right now is clunky and inefficient for most real workflows does that mean its useless? no. I still use it. I think once the context memory problem gets solved this thing is going to be legitimately powerful because the foundation is there but we're not there yet and pretending we are just so you can get engagement is doing a disservice to people who actually want to learn how to use AI properly use it. practice with it. get familiar with it early. but dont throw away the tools that actually work right now just because some guy told you OpenClaw replaces everything soon.