Why, I once encountered a running session longer than a day!

The agent had spun up a backward shell script to watch for the shutdown of another process, but wrote a bug in the script that would have left it running indefinitely until I got home and noticed it.

This was with Fable, no less! And it happened a few more times, though I caught them sooner.

I’m not sure runtime is an important metric at all. Shouldn’t we aim for 0 runtime with maximal results?

How do you get it to spend six hours? I’ve done projects where it would have benefited if it put in extra work.

/goal spend at least 6 hours doing … works

The supervisor agent will keep the session in a loop until 6 hours have passed and eventually the agent will decide to use up the remaining time rather than fighting with it

I wonder how much co2 is being thrown into the atmosphere everyday through the steady stream of “look what I made this LLM do” and endless “benchmarking”?

Sigh

One useful hint here is the API token cost - in this case "about $74 in API tokens" for Opus 5.5.

We know Anthropic run inference at a margin, so that $74 means that the cost of the electricity involved is substantially less than $74. I'd love to know the actual cost there.

I wonder how you only reach $74 in 6 hour runtime. I would have guessed this runtime goes into the hundreds of dollars.

Or I'm realizing that I'm overpaying my tokens...

A lot of that 6 hours isn't spent waiting for the models to return tokens, it's spent executing the code that they've written.

Approximately nothing - the carbon cost of electricity is already very low, and most of the inference costs are GPU and datacenter amortization. You could require all AI inference to be carbon-neutral and that'd barely raise the API prices.

I wonder how much co2 you use to get through the day, and how much you use to remark on other people's creations in a way that suggests you disapprove of their utilization of their available resources. How much co2 do your projects emit, since you seem to be interested in those metrics?

I strongly suspect one HN post (or two now) is not even in the same ballpark. I could ask Claude to figure it out but that would be a complete waste of time and energy. ;)

This has given me a great idea: I have a modest proposal on how to reduce net CO2 emissions while not reducing token consumption at all

USE carbon dioxide to get through the day? What?! We're not plants!

I assure you that if your body did not emit co2 throughout the day you would not continue metabolic processes for an especially long period.

> I wonder how much co2 you use to get through the day

What a terrible thing to say to human! Being (presumably) human, it is their inalienable, natural-born right to do so.

Shame on you.

Intuitively I'd say that just-for-fun AI projects emit orders of magnitude less CO2 than what just-for-fun road trips or air travel emit.

So unless you can show that it approaches all the other frivolous consumption that humans love to waste resources on, maybe we can get back to experimenting with cool technology and talking about it, on this website called "Hacker News"?

Most of that time may have been spent on testing.

The amount of responses to this obvious bot comment is somewhat sad.

claude slop, like from every comments you posted

Couldn't your bot even spare the tokens to read the article? The one-shot broke apart after 90 minutes, not 6 hours, even with explicit instructions.