[tarasyarema]

Hi there!

We've been building https://github.com/desplega-ai/agent-swarm in the open for a while, and one of the things we wanted to do since the start is find how powerful it can be to push for a code mode like env for the swarm.

We tried it and we managed to reduce up +90% of token costs by using swarm scripts (the code mode variant we implemented) for some of our schedules.

This is game changer, as we are able to

1. Run those schedules faster 2. Cheaper 3. More reliably

Been thinking a lot now on how we should prompt and change the default templates to force the agents to build on top of this. If it can be re-used or it's a recurring thing -> scripts.

Thoughts? Have you seen this type of improvements in your setups?

Cheers,

[mpalmer]

    We didn't add this for the post. It's just what the swarm already does, and it's the same machinery behind Script Workflows.

    What we hadn't done is measure it against Anthropic's own yardstick, with our own production data.

You say here that "since the start" you wanted to do a "code mode like env".

But the LLM that wrote your blog post says your system was already doing this and you just hadn't measured it...? How is it that you managed a reduction in token costs by changing nothing?

And can you provide a human-written explanation of the experimental methodology that gets you this miraculous, literally unbelievable 99.2% reduction in tokens?

[tarasyarema]

Hi, the blog post was written assisted with our own swarm, based on the experiment.

Indeed we had the concept of workflows from the start (like n8n), but the point of the scripts is type safe executions that the agents can write, with access to APIs, MCPs, etc.

E.g. this PR from ~10 days ago https://github.com/desplega-ai/agent-swarm/pull/934 was going in this direction, offering a way to defined type safe connections to able to use more the scripts instead of adding MCPs or tools.

e.g. offering a way to do `ctx.api.gmail.searchEmail` coming from an OpenAPI spec.

The whole point of this blog is the "dumb" realisation of the following:

If you have an agent (say Claude Tag, or some other fancy Slack bot) that you keep asking the same things (e.g. check DataDog error and correlate PRs on a timeframe) and the agent does that process agentically, you are wasting tokens.

Now, if you give a way for the agents to write these scripts (code mode) the next time a task like that happens then essentially you just run that and you get easily 90% cost reduction in THAT process. Not globally, obviously.

[moberemk]

No offence but, “if you find yourself doing the same thing over and over again write a script for it” isn’t exactly novel thinking in the software field…this was the same process when it was a human doing it, we just measure that in time instead of tokens

[tarasyarema]

yes, 100%

but it's not that simple. the whole point of this is that if you have an AI system that is in charge of this, what you need is a way for it to build those deterministic software based parts, while maintaining the AI routing and reasoning.

with time, our internal swarm has been building those "internal" software that is tedious and boring to build, where it makes sense.

e.g. we also have a "pages" thing (like claude arifacts) that it uses to join scripts, html and other components to build dashboards for us

[lyall]

There might be something valuable in this article, but if the you couldn't be bothered to write it yourself, why should we be bothered to read it?

[tarasyarema]

This is based on a multi-day convo I had with our swarm. I might have not wrote it myself, but it's based on human interaction and thought process.

Any recommendations on how I could make it better next time, so you do not feel like wasting your time (which I get and feel, dont get me wrong!)

[subscribed]

Indeed it's the HN variant of "I asked ChatGPT and it said...."

I wish there was a way to tag this type of the posts, but at least my HN client allows to tag the authors.

[neongreen]

Look at https://randomlabs.ai/blog (Slate), they're exactly that and you can play with it

[rmsaksida]

I'm sure there is some valuable information in the article, but the LLM-generated prose is unreadable.