Elektrine lite

← Feed

@revk@toot.me.uk

2026-09-19 16:33 UTC

FFS people poking AI at my circuit designs now. May be time to retire - I don’t want this shit. And it was crap... Saying 196mA charge rate is close to the 200mA limit. Yeh, well, and, well, FFS. Saying a GPIO has no pull up, so may not read properly until initialised with internal pull up - well Duh! So I'll initialise before polling it, FFS. Why even post this crap to me?!?!? I need to instant ban/block any AI crap!

Replies (10)

  • @Miq@hessen.social 2026-09-19 17:03

    @revk@toot.me.uk I feel you. Block and ban seems to be the measure in order. 😡

    Open ##4766057

  • @revk@toot.me.uk 2026-09-19 16:49

    He reopened the issue he posted AI crap to, and not even a comment why. Had it been an apology or explanation, fine, but nothing. Now actually blocked. Arse.

    Open ##5012715

  • @revk@toot.me.uk 2026-09-19 16:54

    I’m quite cross on both points. Being just under max charge rate seems “ideal”. And using GPIO internal pull up instead of external is what the fuck it is there for. This is total and utter bullshit. And actual waste of my time. Bad enough if/when AI finds a mistake, but finding things I clearly did right and calling me out for it is just fucking stupid.

    Open ##5012716

  • @miblo@mas.to 2026-09-19 16:49

    @revk@toot.me.uk *nods appreciatively* My mind's been turning to activities centred around our senses of taste and smell, thinking surely they'll be safe from the LLM'ings for the foreseeable. Either that or we'll finally get smell-o-vision…

    Open ##5012732

  • @sconient@techhub.social 2026-09-19 19:14

    @revk@toot.me.uk As long as you don't care what the pin state is during the startup phase, when there's no microcontroller power, or if you don't need a strong pullup/down for a particular reason, then just relying on the internal pullup/down is totally fine. As you said, that's why it's there. One of the (many) big issues I see with AI answers is that they are presented with so much authority, when in reality they are no better than random answers that would otherwise have been found with a web search. Ask any (decent) engineer and you'll get an appropriate answer with a list of caveats and things to watch out for, and maybe something will be learned in the process. A good engineer knows when to say "I don't know" as well, instead of just giving a confidently wrong answer. (Obviously this applies to any profession, but being an engineer myself that's the majority of interactions I experience.)

    Open ##5012736

  • @benjamineskola@hachyderm.io 2026-09-20 06:39

    @revk@toot.me.uk I don’t know anything about electronics but this lines up with what I’ve seen of LLM-extruded ’code reviews’. Stuff that might be technically correct but irrelevant at best, mixed with stuff that’s just outright wrong.

    Open ##5012740

  • @revk@toot.me.uk please don't let them kill your creativity. Block and move along. Maybe write a agents.md demanding money for reports lol

    Open ##5012741

  • @GerardThornley@hachyderm.io 2026-09-20 09:29

    @revk@toot.me.uk "... a GPIO has no pull up, so may not read properly until initialised with internal pull up" so... "You haven't added included an unnecessary redundant component." Yikes, how awful!

    Open ##5012742

  • @dr2chase@ohai.social 2026-09-20 13:06

    @revk@toot.me.uk been lucky enough to avoid this, but cannot help but imagine either the missed details if the LLM did not realize the circuit was intended for a bicycle (vibration, damp, static, wide temperature range) or the irrelevant pickiness if it applied bicycle environmental rules to an indoor circuit.

    Open ##5012743

  • @Stubbs@mastodon.me.uk 2026-09-20 16:17

    @revk@toot.me.uk Do you reckon it's people who have triggered it or bots that have triggered and submitted it?

    Open ##5012745