Post

If they were trained to help find/setup shortcuts and then not get used anymore for that task they could be a net positive (for this one thing) but of course we can't have that, and all the other issues ensure it's an overall net negative either way.

One of their examples was actually that they had it check if their dehumidifier needed to be empty which I have previously built as a worker (with no AI) github.com/willmanduffy...

My partner had a work meeting yesterday in which word came down from on-high 'NO USING OPENCLAW FOR ANYTHING' don't know if folks had been or what, the vibe he got was that management wanted to get out ahead of it

Yes, but now you can do it without having to learn anything, and instead of the pain and boredom of learning, you get the unpredictable reward mechanics of a slot machine! (Also, the slot machine will shoot you in the face occasionally.)

First adopter mindset is one hell of a drug. You’re right, every piece of software would be better if it had tons of features and capabilities it doesn’t currently have.

And LLMs are really good at helping you build deterministic software and automation workflows if you know what you're doing.

the OTA model is underrated for hardware. builders ship fast, builders fix fast.

But those local deterministic tools don't siphon micro payments to billionaires every time they achieve a trivial task. Won't somebody think of the billionaires?

The BASIC snippet at the end is chef's kiss. "Garbage in, garbage out" has been a known principle since the 1950s but somehow every hype cycle rediscovers it fresh.

Can't get anyone to write documentation for an API, but suddenly when they need to write a clear and comprehensive prompt for an MCP...

But you don't get it! I bought 4 Mac Minis and installed OpenClaw to orchestrate & fully automate my email organization, and now I've achieved Inbox Zero* *because OpenClaw deleted all of my emails and my entire email account

This is just about everything I’ve seen demoed. Things that could easily be solved with existing technology more reliably and most importantly, more accurately. If the job requires precision, AI is not the solution.