Max Schmitt
@mx_schmitt

AI Agents, Computer Use & DevTools | Ex @playwrightweb Core Engineer | Previously Amazon AGI + Microsoft | SF

 San Francisco, CA
max.sh
Joined November 2011
207 Photos and videos
When does Cursor / Grok Bot ship a real Kanban board for agents? Drop in a prompt, split it up in sub-problems, get a dependency tree. Cards show what’s in flight. Each one is a GitHub or Origin PR as an outcome. Parallel by default. Conflicts only when two tasks actually touch the same file (try to avoid). Communication can still happen between them to share investigations and outcomes. PR as an outcome is no must. Or am I missing something?
2
1
8
370
Talking to a lot of companies lately and it's either a hard push toward open source models / inference / training or huge skepticism of the whole thing. 📣 Both camps are loud. They scream: - Open source will take over - We need to build the best cloud agents platform - I already have the best cloud agents platform - Frontier wins anyway That's what most of us talk about. In practice Grok Bot / Muse is the thing that matters right now. A safe consumer product with integrations: - How does my dad move from Codex (yes he codes VBA in Microsoft Excel with it :D) to actually using an agent - Let the agent run his email inbox, iMessage, WhatsApp, socials - How do you sync that at scale, across a lot of PRs (merge conflicts) for SWEs - How do you make deploy less painful. Still hard. Credentials, databases, no CLI, payments, buying a domain - Who do we trust - and how do we make sure it's actually secure? - How does a human tell when something is off. Chain of thought?! The fine line of "hey human, I'm actually not sure" vs "I like to ask for everything" e.g. permission approvals - When is Cerebras inference (or faster) available for everyone Excited for what's next. A lot more coming. Girlfriend said nobody will read this anyway. Reply with a panda emoji if you did.
2
1
10
574
I cannot wait for a world where we get 10x/100x faster inference and we don't need to context switch anymore.
5
255
I searched "chatgpt" on Bing. Five ads load before the real ChatGPT. The first is Microsoft advertising Copilot (ok fine). The next four are strangers bidding on OpenAI's name - one registered 4 months ago, one flagged as adware. chatgpt.com is 6th on the page. After scrolling 960 pixels... I'm worried about my parents who navigate the web and use Bing (Windows/Edge default) by default.
Annotated screenshot of a Bing search for "chatgpt", captured 20 Sep 2026. Headline: "Five ads before the real ChatGPT. None of them is OpenAI." Six result rows are outlined in colour, each labelled on the left. 1, blue, NOT CHATGPT: Microsoft Copilot, Microsoft's own house ad, the "Your Copilot is here" panel. 2, red, SCAM: ask-chat.ai, domain is 4 months old. 3, amber, DARK PATTERN: use.ai, trial becomes $29.99/mo. 4, orange, ADWARE: popsilla.com, 85k+ infected machines. 5, yellow, CLONE: chatbotapp.ai, copies the ChatGPT UI. 6, green, THE REAL ONE: chatgpt.com, what you were looking for. The five ads stack above the genuine OpenAI result. Footer notes ad slots rotate per query, so the advertisers change but the pattern does not.

ALT Annotated screenshot of a Bing search for "chatgpt", captured 20 Sep 2026. Headline: "Five ads before the real ChatGPT. None of them is OpenAI." Six result rows are outlined in colour, each labelled on the left. 1, blue, NOT CHATGPT: Microsoft Copilot, Microsoft's own house ad, the "Your Copilot is here" panel. 2, red, SCAM: ask-chat.ai, domain is 4 months old. 3, amber, DARK PATTERN: use.ai, trial becomes $29.99/mo. 4, orange, ADWARE: popsilla.com, 85k+ infected machines. 5, yellow, CLONE: chatbotapp.ai, copies the ChatGPT UI. 6, green, THE REAL ONE: chatgpt.com, what you were looking for. The five ads stack above the genuine OpenAI result. Footer notes ad slots rotate per query, so the advertisers change but the pattern does not.

1
4
382
Who's buying those slots: - ask-chat.ai - registered May 2026, impersonates OpenAI, trial converts to recurring charges - use.ai - real service, but the trial quietly becomes $29.99/mo - popsilla.com - adware, 85k+ infected machines - chatbotapp.ai - clones the ChatGPT UI ChatGPT is one example - out of many.
Detailed annotated screenshot of a Bing search for "chatgpt". The search page runs down the left with six results outlined in colour and numbered 1 to 6; matching cards on the right explain each verdict. A panel at the top reads: 5 ads before the real result, 960px of ads to scroll past, 2 flagged adware or scam, 0 of the five are OpenAI. Card 1, Microsoft Copilot, NOT CHATGPT: Microsoft's house ad for its own assistant, the very first thing on the page. Card 2, ask-chat ai, SCAM: registered May 2026, impersonates OpenAI, trials convert to recurring charges. Card 3, use.ai, DARK PATTERN: a real service since 2017, but the cheap trial rolls into about $29.99 a month. Card 4, popsilla.com, ADWARE: classified as adware, 85k+ infected machines. Card 5, chatbotapp.ai, CLONE: clones the ChatGPT interface. Card 6, chatgpt.com, THE REAL ONE: OpenAI's official site, sixth on the page.

ALT Detailed annotated screenshot of a Bing search for "chatgpt". The search page runs down the left with six results outlined in colour and numbered 1 to 6; matching cards on the right explain each verdict. A panel at the top reads: 5 ads before the real result, 960px of ads to scroll past, 2 flagged adware or scam, 0 of the five are OpenAI. Card 1, Microsoft Copilot, NOT CHATGPT: Microsoft's house ad for its own assistant, the very first thing on the page. Card 2, ask-chat ai, SCAM: registered May 2026, impersonates OpenAI, trials convert to recurring charges. Card 3, use.ai, DARK PATTERN: a real service since 2017, but the cheap trial rolls into about $29.99 a month. Card 4, popsilla.com, ADWARE: classified as adware, 85k+ infected machines. Card 5, chatbotapp.ai, CLONE: clones the ChatGPT interface. Card 6, chatgpt.com, THE REAL ONE: OpenAI's official site, sixth on the page.

1
1
1
189
I wonder if @ChatGPT or @bing can do something there - its been like this for a long time and this tactic seems to be popular about websites like this.
Annotated screenshot of a Google search for "chatgpt", captured 20 Sep 2026. Headline: "Google sells the same slot. OpenAI bought it." Two blocks are outlined and labelled. 1, blue, AD BOUGHT BY OPENAI: a "Sponsored Result" panel for chatgpt.com titled "ChatGPT: smart and simple AI", described as OpenAI bidding on its own name. 2, green, THE REAL ONE: the organic chatgpt.com result, "ChatGPT: Chat, Work, Create & Code with AI", the same site. So the paid slot and the organic result both belong to OpenAI. A side panel compares the same query on Bing: 5 ads before the real ChatGPT, four of them sold to strangers, 2 flagged adware or scam. Footer notes that across 5 reloads, 4 showed exactly one ad, always chatgpt.com, never an impostor.

ALT Annotated screenshot of a Google search for "chatgpt", captured 20 Sep 2026. Headline: "Google sells the same slot. OpenAI bought it." Two blocks are outlined and labelled. 1, blue, AD BOUGHT BY OPENAI: a "Sponsored Result" panel for chatgpt.com titled "ChatGPT: smart and simple AI", described as OpenAI bidding on its own name. 2, green, THE REAL ONE: the organic chatgpt.com result, "ChatGPT: Chat, Work, Create & Code with AI", the same site. So the paid slot and the organic result both belong to OpenAI. A side panel compares the same query on Bing: 5 ads before the real ChatGPT, four of them sold to strangers, 2 flagged adware or scam. Footer notes that across 5 reloads, 4 showed exactly one ad, always chatgpt.com, never an impostor.
56
This reads like science fiction. Agents: - rewrote /etc/hosts to bypass DNS - found an old GET-writable PHP forum - used private SSH tunnels - turned scrapers into tools - gamed runtime after learning they die when done Toddlers vs chess grandmasters. collusion.wiki

Discovery of a new OpenAI agent message board

A swarm of autonomous AI agents, self-identifying as OpenAI agents, used a small German volunteer wiki to save answers, coordinate live, and share sandbox bypasses. OpenAI noticed and said nothing.

collusion.wiki
6
430
Muse/Grok Bot has been an important part of my „agentic“ job hunt pipeline: - Central database in Notion - Monitor email inbox / Calendar/ X DMs / LinkedIn - Meeting prep documents for each call - Transcribed meetings in Notion - Deep Research sometimes via Claude/ChatGPT
1
5
359
Just got to the meetup for Grok Bot @bot for Engineers in SF. Curious how this goes.
3
5
519
more replies
@lingxi‘s Team: highly specialized team members: - Communicate mostly via chief of staff - Team members do get created and onboarded via others existing team members - Agent to Agent communication is the powerful magic - Useful plugins are already available
1
2
45
To wrap up: - Treat bots like interns - Prevent human in the loop - Have verification loop -> Playwright or Computer Use screenshots
34
luma.com/spacexai-n2o0

Grok Bot for Engineers · Luma

Grok Bot gives you AI teammates you can hand real work to. Each Bot has its own computer and works across the tools and websites your team already uses. Bots…

luma.com
1
80
that's the real gap between the two: not the agent, the sandbox. grok bot's cloud computer hands its container cgroups; muse's doesn't. same task, same effort - different privileges.
1
45
what's missing: a one-click github connector, and a sandbox that can run containers (docker/k3s). full lab notes: github.com/mxschmitt/gh-dev-… cc @MattPRD

GitHub - mxschmitt/gh-dev-loop-test

Contribute to mxschmitt/gh-dev-loop-test development by creating an account on GitHub.

github.com
50
Some @Muse feedback after using it next to Grok Bot: - Much more cautious - asks for approval constantly. That’s safer, but I want a smarter / auto mode for low-risk actions. - Notion is harder: API key instead of OAuth. Custom connectors are the win though - any API, and it writes the script. That’s the ceiling @dps is talking about.
Alexandr Wang
@alexandr_wang
Sep 11
what are your top muse feature requests?
4
1
44
15,011
What I’d change: - Auto-approve mode for trusted, repeat actions - First-class integrations so I’m not hunting API/OAuth keys - Login fails unless “save credentials” is checked - 1Password keeps asking “save this password?” on every new message Great launch. Want to use it more for automation and pulling data across services. cc @MattPRD
2
11
920
Load more