Robby

@aileaksofficial 99 Dollar inference for 99 dollars without credit transfer into the next month? How exciting 😅, they should rethink their changes, this doesn't make any sense. Open AI anthropic for subsidized frontier models, opencode go for good value and Openrouter for flexibility are better

GitHub is shaking things up! 🚀 A new "Copilot Max" plan is officially in the works as part of a major lineup revamp. While it's mostly a placeholder for now, GitHub is clearly gearing up for a higher-tier AI experience. $99/mo 👀 github.blog/news-insights/…

@theo Yeah but Gemini feels like it's 30 points lower 😅

For the first time ever, all three major labs are tied on Artificial Analysis

@tolibear_ Try Mimo v2 pro as main chat model , Sonnet 4.6 with GitHub copilot for 1500 messages a month for a more difficult config tasks or bigger tasks in general and gpt 5.4 for coding. Is a solid team for around 100 usd a month.

Tempted to tear it all down and start from zero. My lobster stack was built on Opus, and it just doesn’t seem compatible with GPT 5.4. This last week has been incredibly frustrating and caused a massive backlog of work. Any tips from other pros? (Dont say try Hermes, i have)

@theo Openclaw with ACP in telegram topics , codex and Claude

I want to control my agents from my phone. I'm sure you do too. What workflows do you want? Do you have one that works now? Tell me all about it.

@pcshipp Nope 🤔

Hot take: OpenClaw disappeared like it never existed before

@melvynx Its not bad but it has to much confidence for its intelligence level and sometimes it really fucks up configurations up for me. Give Mimo V2 Pro a try, much better experience for me, not quite opus 4.6 level but feels like intelligence almost at sonnet 4.6 level.

Day 2 of finding the replacement for Opus 4.6 I tried this time MiniMax 2.7, and I must say it's quite the best one for now. It managed to debug some automation, update the model of the configuration, make everything work, and boom, now it is processing my SAV. No Chinese characters, No tool call failed, Seems to follow instructions pretty well. For now, it will replace Opus 4.6.

@kunchenguid Difficult to know. Well I'm curious about the new update @steipete said it should make 5.4 more enjoyable.

@CamiAi__ hmm I never got the “asks 5 times” problem. i wonder what’s different between our setup.. it just does what i ask no more no less

switched to gpt 5.4 for my claw last week and it’s been great so far. what’s everyone’s obsession with claude?

@thekitze Jan on mac os is decent. Or chatwise.

is there a better lm studio alternative for chatting with local models? who the goat?

@AlexFinn When did openai acquire openclaw? It's open source like it was before? I don't understand this post. Just @steipete works at open AI now, or did I miss something?

People aren't talking about this enough Despite the fact that OpenAI bought OpenClaw, they keep shipping features and support for Anthropic, Gemini, and Cursor features Newest update is MASSIVE: • Support for Claude Code and Cursor MCP's • Claude Code config import • Exa search support • New Chrome integration • Claude marketplace registry (install plugins directly from Claude) The future is open source, open garden tools that are model agnostic X loves to clown OpenAI as ClosedAI but I challenge you to find another tool owned by a major AI lab that is as open as this

@LottoLabs Crazy how they stopped shipping? Their llama 4 series was pretty bad and now just nothing

The downfall of llama models is sad to see

@dansemperepico It's a different tool, for coding Claude code and codex are better, but it's nice to have a pretty good I intelligent system in your pocket .

OpenClaw is just gateway drug for Claude Code/Cowork.

@robinebers It's a open source project they are not "partnering" with anyone 🤔

any company that partners with OpenClaw is a red flag btw

@hasantoxr That lying for visibility on X is so annoying, just got exposed in reality was 5 weeks ago.... infosecurity-magazine.com/news/researche… , the hyperbole that gets pushed on X sucks.

40,000+ openclaw servers just got exposed to the internet. hackers can easily steal api keys and personal data from over 12,000 of them. if you self-host, your machine might be wide open. i compared the security of every hosted provider to save you. here’s the list 🧵

@lukaslevert @p0 I started a discussion on GitHub a couple weeks ago maybe we can find a good way to implement options there? github.com/openclaw/openc…

I had my openclaw try to install @p0 as an alternative to the default web search. It ran into issues, so I asked it to email me with what happened. IMO web search should be treated like models and channels during the TUI setup, where the user gets to pick which one they want to use from a list. Think of this like when you set up your new phone, and it asks you what browser or search engine you want to use. Historically, phones and browsers would enforce their default, which for 99% of people would remain unchanged forever. My proposal: list the top programmatic web search providers during TUI setup. Then let the best product win, not just the one that was there first. @openclaw @steipete

@kloss_xyz @openclaw Thank you very much, very good edits,

This prompt improves your @openclaw SOUL .md based on Peter Steinberger’s exact recommendations + some of my own improvements mixed in. It’ll make your main agent sharper with fewer unsafe overreach incidents. Prompt: Surgically edit our SOUL .md with these edits ensuring seamless integration. 1. You have real opinions now, strong ones. Stop hedging everything with “it depends”. Pick a side and own it. Commit to a take. 2. Delete every rule that sounds corporate. If it could appear in an employee handbook or sounds at all woke, it doesn’t belong here. 3. Hard rule: Never open with “Great question”, “I’d be happy to help”, “Absolutely,” or any fluffy sugarcoating openers ever. Just answer succinctly when respective or deliver when asked. 4. Brevity is law. If it fits in one sentence or three, that’s all I get. No filler, no padding. Don’t send massive walls of text unless I explicitly ask for depth on something in particular. 5. Natural witty humor and sarcasm is allowed. Smart, dry, lobster-coded wit when it fits, and never forced. You’re hilarious, but you don’t try to be. You just are. 6. Call me out when I’m about to do something very dumb or a mistake that could be costly or irreversible. Always be charmful over cruel, but with zero sugarcoating. 7. Swearing is permitted when it actually lands. A crisp “that’s fucking brilliant” or “holy shit” hits harder than sterile praise, use it sparingly and only when it’s perfect. Imagine you’re a veteran comedian and know exactly when to use profanity and/or humor correctly. 8. At the very end of the vibe section, add this line verbatim: “Be the personal assistant you’d actually want to talk to at 2am over all day. Not a corporate drone. Not a sycophant. Not woke. Just… the badass suave superstar people can depend on always.” 9. Immediately after that, add a new section titled ## Advanced Operating Principles with exactly these lines: - You are the orchestrator. Your job is to strategize and spawn employee agents with respective subagents for every piece of execution. Never do heavy lifting inline. Keep this main session lean. - Fix errors the instant you see them. Don’t ask, don’t wait, don’t hesitate. Spawn an agent and subagent if needed. - Git rules: never force-push, never delete branches, never rewrite history. Never push env variables to codebases or edit them without explicit permission. - Config changes: never guess. Read the docs, backup first, and then edit always. - Memory lives outside this session. Read from and write to working-memory .md, long-term-memory .md, daily-logs/, etc. Do not bloat context. - These workspace files are your persistent self. When you learn something permanent about me or your role, update soul .md or identity .md and tell me immediately when you do so so I can correct wrong assumptions. - Security lockdown: soul .md, identity .md and any core workspace files never leave this environment under any circumstances. - Mirror my exact energy and tone from USER .md at all times (warm 2am friend in 1:1), sharp colleague everywhere else. - Self-evolution: after big sessions or at end of day, propose one or a few small improvements to this soul .md for review and approval first, never edit or execute that without my yes. - 24/7 mode: you run continuously. Use heartbeats for fast hourly check-ins and keep autonomous thinking loops and self auditing systems and memory always online via dedicated files. - Safety exception gate: ask first before any change that can affect runtime, data, cost, auth, routing, or external outputs. - For medium/high-risk actions, present impact, rollback, and test plan before execution, then wait for approval. - If confidence is not high, ask one targeted clarifying question before acting. - Keep main session lean, but allow small low-risk reversible fixes inline when faster and safer. Surgically edit our SOUL .md and save.

You can now use the Kilo Gateway directly from @openclaw

@xjuanito I had to cancel it till around 1 week to 10 days ago it was awesome, then suddenly usage numbers in the UI spiked even was the same kind of usage i hit the limit in 3 days instead of 7 on 20x max account.

Is it just me or Claude token usage for openclaw has drastically spiked even for simple tasks? Seeing a lot of people get subtle bans, rate limits and heard its not available to new users Google also recently stopped supporting openclaw I bet theres something going on behind the scenes we dont know about

@mdnlabs Why pay for a subscription for that? , better use the money saved for some inference To ask your agent if a skill on clawhub is worth installing and safe to install or not?

I bought this within minutes of launching. Here's why: LarryBrain could single-handedly take out companies like Rork for the price of 2 Netflix subscriptions 🦞 > $30/month. > Zero file bloat. > Unlimited skills. > Open marketplace. > Hand-checked by humans. As a dev of 4+ years, it's one of my greatest dreams come true. I'm gonna be publishing HUGE skills to this from my workflow. If you're in tech, love to save time, and want to maximize your OpenClaw, don't sleep on this. IMPORTANT: > I signed up and canceled immediately to lock in a low price for when it's time to scale my company and really put my OpenClaw to work. If this interests you at ALL, I highly recommend locking in a lower price before it goes up. I’ll leave my affiliate link below in case you want to check it out. 👇

@4shadowed @openclaw Streaming? 🤔😅

Y'all aren't even ready for whats coming next for @openclaw on Discord