603
"lessons learned" (lemmy.ml)
submitted 9 months ago by to c/fuck_ai@lemmy.world
146 comments fedilink hide all child comments
 

screenshot of followup to the first tweet, also by @BenjaminDEKR: " Made some adjustments based on lessons learned. Combined: roughly 200-400x cheaper heartbeat operation."

https://xcancel.com/BenjaminDEKR/status/2017660150463582282

via this bluesky post:

screenshot of bluesky post by @rusty.todayintabs.com with text "I’m starting to think the people who are excited about “AI agents” have literally never used a computer in their lives" and screenshots of the above tweets

screenshot of tweet by @BenjaminDEKR with text: OpenClaw is interesting, but will also drain your wallet if you aren't careful. Last night around midnight I loaded my Anthropic API account with $20, then went to bed. When I woke up, my Anthropic balance was $0. Opus was checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night." Doing literally nothing, OpenClaw spent the entire balance. How? The "Heartbeat" cron job, even though literally the only thing I had going was one silly reminder, ("remind me tomorrow to get milk") 1. Sent ~120,000 tokens of context to Opus 4.5 2. Opus read HEARTBEAT .md, thought about reminders 3. Replied "HEARTBEAT_OK" 4. Cost: ~$0.75 per heartbeat (cache writes) The damage: - Overnight = ~25+ heartbeats - 25 × $0.75 = ~$18.75 just from heartbeats alone - Plus regular conversation = ~$20 total The absurdity: Opus was essentially checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night." The problem is: 1. Heartbeat uses Opus (most expensive model) for a trivial check 2. Sends the entire conversation context (~120k tokens) each time 3. Runs every 30 minutes regardless of whether anything needs checking That's $750 a month if this runs, to occasionally remind me stuff? Yeah, no. Not great.
you are viewing a single comment's thread
view the rest of the comments
[–] 54 points 9 months ago (5 children)

What an unhinged thing to rely on an llm for.

Had a Cron job running

So they set up a Cron job to ask an llm to remind them of something that the cron job itself could have just reminded them?

Everything about this is so wrong lol

  • permalink
  • fedilink
  • source
  • hideshow 5 child comments
    •
  • [–] 24 points 9 months ago

    But a simple reminder is so... impersonable...

    Why settle for a basic reminder when you could have a personal AI agentic assistant tailored toward your needs to remind you things like "buy milk" or "wash your hands after going to the bathroom"?

    Your personal AI agentic assistant gets to know you on a deeper level, learning your preferences and adjusting to your unique needs, so that it feels more like having a loyally devoted butler to gently wake you and remind you to wipe your ass.

    Some day, as the technology matures, maybe your personal AI agentic assistant will even wipe your ass for you, or buy your milk. But only after washing its hands in between!

    -him, probably

  • permalink
  • fedilink
  • source
  • parent
  • [–] 11 points 9 months ago* (2 children)

    It only makes sense if it was checking for it being daytime (i.e. after sunrise and before sunset) which you cannot do in cron, rather than check for a specific hour.

    Even then, using an LLM is about the stupidest way imaginable to do it since it's not as if "when is it sunrise/sunset at a specific latitude and longitude and day of the year" can't just be calculated with a formula or looked up in a table of values - its not as if the sunrise and sunset hours given latitude, longitude and day of the year change from year to year.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 8 points 9 months ago (1 child)

    It's just astonishing how many thoroughly solved software problems are now being delegated to LLMs.

    It shows a fundamental misunderstanding/delusion about what an LLM actually is.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •