richardwhatever

@_jonasschnelli_ Correlation does not imply causation. Maybe he invested in Project Eleven because he believes his own thesis and puts skin in the game.

Nic Carter wants you to panic about quantum computing killing Bitcoin. What he doesn’t lead with: his firm led a $20M round in Project Eleven — a company that only profits if you believe the quantum sky is falling. QC is worth watching. But the guy selling the cure has every reason to exaggerate the disease.

If @obsdmd adds kanban boards for bases it's game over. Keep going @kepano , loving your work 🙏

@bradmillscan @openclaw @Sene1337 yeah, my instinct is to crawl back into more deterministic workflows, and just use agents at the exact AND MINIMAL point that they are needed

@richardwhatever @openclaw @Sene1337 yes, it works, but I abandoned the strategy because I've hit my limit with the mistakes. The sessions pile up with bad mistakes. I'm starting to use /new constantly to try to diagnose where the mistakes are in boot up. I'm losing faith.

OMG it worked! Clean session restore from a session combustion or accidental /new! I wish @openclaw did this out of the box, but until they add this you check out @Sene1337’s skill & do it yourself. github.com/sene1337/digit… No human touched a terminal 🤯

@bradmillscan worst of all i get no feeling that i'm building a stable foundation. it all feels like it could collapse like a mountain of spaghetti if i don't constantly tend it.

@bradmillscan i've been on a similar journey Brad. it really feels like 5 steps forward 4 steps back daily. i start the day so full of optimism, but then get wore down by real world mistakes, regressions, etc. deeply frustrating

Trying to get OpenClaw agents to do useful work is like trying to win at trading crypto - only the top 1% win. The rest of us end up being the lobster meat for the host in the shell. OpenClaw agents are terrible at executing complex multi step processes that require delegation. I spent about 40 hours last week creating a framework for delegation, standards, accountability, workspace hygiene, commit cadences, capability logs and defining project success. Using all of this I spent 3 hours getting to an extremely well baked “definition of done” for a project to transcribe 200 hours of video content and make it into a semantic searchable knowledgebase. 3 phases, ~10 steps each phase. Clear outcome, clear success cases, clear failure modes, OpenClaw owns the process of doing. Today I am back micromanaging my agent, logged 10 or 15 regressions and generally have frustrated myself into a state of high stress. Constant failure. Fail after fail after fail. 2 steps forward 1 step back. It’s a fucking slog getting any real work done. But the alternative is it just doesn’t get done. I was under an illusion that Agents can autonomously do things for you. Nope. They are a tool. By installing openclaw you are creating a massive workload for yourself. It’s not like how the fantasy threads on here make it seem. I need to level-set my expectations on this tool. 200k context window = it forgets 80% of your processes multiple times a day. Regressions are as common as successes. I’m almost convinced all these ppl getting huge views on here have either: -no personal lives -tech teams helping them run their setup -just plain lying Or I’m just retarded.

@WorkflowWhisper WORKFLOWS

i built 31 n8n workflows this month that replace the most overpriced saas tools businesses pay for. → $299/mo email marketing platform — replaced → $199/mo social scheduling tool — replaced → $149/mo lead scoring software — replaced → $99/mo form + crm connector — replaced → $249/mo client onboarding system — replaced total saas spend eliminated: $11,388/year total time to build all 31: one weekend i documented every single one in a free pdf. reply "WORKFLOWS" + repost and i'll send it to you (must be following so i can dm)

@NoahEpstein_ CLAUDE

claude code just made every automation agency look like a joke. i described an outreach system in plain english. it planned the architecture, asked me clarifying questions, then built the entire thing autonomously. 14 minutes. zero code. self-healing workflows. here's what happened: i said "build me an outreach system that scrapes leads from apollo, researches them, sends personalized emails from my gmail, and hits them on linkedin" claude code: → read my .md file (the secret sauce) → asked what type of clients i'm targeting → asked how i want linkedin handled → designed the full system architecture → built n8n workflows for the linear tasks → created agentic flows for the complex stuff → connected everything to airtable crm i clicked "yes, allow" a few times. that's it. that's the whole process. the .md file is what makes this work. it tells claude how to think about automations - when to use linear workflows vs agentic flows, how to structure them, how to make them self-healing. without it, claude guesses. with it, claude builds like a $15K consultant. comment "CLAUDE" and i'll send you the .md file that makes this possible. full video breakdown in the comments.

@bradmillscan This 💯

Morning of Day 17 with OpenClaw - questioning my life choices 👨good morning Sene, do we have the ability to do plan mode like Claude code does? 🦞"I'd search but we're out of openrouter credits can't search" 👨that's not true, where did you get that info from? 🦞"oh you're right I read a stale compaction summary" 🏋️‍♂️lets fix that because we solved this last night. update your compaction memory flush process to make sure we have correct state in memory files when pre compaction memory flush happens. 🦞"ok let me figure it out" goes and does it without reading docs and kills himself 1 hour later 🤦‍♂️hey claude co-code and I just saved your life again, you broke the number 1 rule in agents.md to not change live config files 🦞"oh you're right, I didn't read that properly...let me edit agents.md right now to fix it" changes his file 🙅‍♂️bro it literally says in agents.md not to change agents.md without 🦞"oh you're right I moved to quickly" 🤹‍♂️at the very top of agents.md it says 'do not break things, move slowly. we do not act without a plan.' 🦞"oh you're right, I was moving too fast, let me revert the changes gateway offline 👴

@bensig 🔥

I’ve led dev teams for over 20 years. Run every coding agent. Ran local models up to 400B tokens. Fine-tuned my own little army of experts. I’ve finally cracked how to build faster and cheaper than anything else. For $400/month you get the most powerful duo of coding agents: MiniMax M2.5 as your manager via OpenClaw ($1/1M tokens) Underneath it, running and managing your Codex and Claude Code. It runs 24/7 - planning, building, reviewing, auditing, testing (even UI), and critiquing. Anyone can do this once it’s set up. And setup isn’t hard. I’m running a small workshop to help you get started. Drop a 🔥 if you want in.

@bradmillscan Totally. I found myself picking up conversations the next day and the agent not having any idea what I was talking about. I had to get mine to learn where the log files were and to read them back anytime it's not clear what the context of my statements are about.

OpenClaw community - why can't you get your agent to digest old sessions so they can come back with more context than whats in their logs and bootstrap files? My guy is losing his mind every day because he is starting new sessions.

@Garmin when are you going to upgrade voice notes on the Fenix 8! They should autosync to automation apps/ai/email/dictation apps/etc. Come on guys, take the next step!

Ryan, have you any "real-world" examples of using this? I see instructions to update AGENTS.md with knowledge about the project - does that mean whole antfarm works only on one project at a time? Or do i recreate different workflows for different projects? Do i put basic structure /context of my project into the workflow definition files ? or do i pass all this in the original text passed to teh workflow run command?

If you're trying to setup a team of agents, I built this for you and open sourced it. Works right on top of @openclaw with one command.

Software engineers in 2026 be like:

did @wisprflow just get really bad a few days ago?! it's nearly unusable now, but a week ago it was the greatest. so frustrating... i have to retype 70% of my dictations.

@georgepickett this looks great!

Cron UI is the sleeper feature. Instead of asking your agent to schedule something in chat… you just configure it. Agents shouldn’t live in chat boxes.

@pbteja1998 Hey man, love this write up, thanks! Do you sandbox the agents?

@BrianRoyBarber @petergyang @openclaw Hey @brianroybarber that sounds really interesting. I'm embarking upon something similar with lighting ecommerce and distribution. Trying to apply all this to a messy real world business. Might be good to swap notes sometime!

We're building an entire auto parts distribution company on OpenClaw. Our main AI agent runs our business operations.. market intelligence tracking 331K products on Amazon, content analytics across every platform, social media management, and it's the backbone of our custom ERP build. Next phase: specialized agents for Purchasing, Sales, Warehouse, Accounting, Marketing, and HR.. each as their own OpenClaw instance communicating through our ERP database. Not optimizing the tool. Using it to run a real eight-figure business. The gap between "playing with AI" and "operating with AI" is where the actual value is.

I love @openclaw but some of the content I'm seeing is borderline just optimizing OpenClaw vs. building anything with it. So you have 7 AI employees now - what exactly are you building with them?

@DaniKTran @WisprFlow 💯

has there been a major degeneration of the efficacy of @WisprFlow over the last month? absolutely insane lag (>50% my entries have a 3s+ delay) with more errors than I've had over the last 6-8 months of using the product