@voidisomorphism @theo both up as PRs, both still open.
hosted ui: one star ask after the 5th prompt
agent-fs: same after the 5th write
close it once and it never comes back
agent-swarm#2652
agent-fs#147
2 likes · 6 views
check -> agent-swarm.dev (OSS/MIT) / repo: desplega-ai/agent-swarm
It's exactly that.
The key point is -> dynamic workflows, or static ones should contain guardrails and scripts enforcing certain rules. Check the code!
It's non-trivial for sure
I do think apps like these are interesting. My question is what exactly you need from humans as input on what scale to get big benefits from it?
desplega-ai/agent-swarm
1 like
Ask HN: How do you decide to cut scope when AI makes it so easy to add?
I've been having this conversation with my co-founder for months.
It's so easy to add incremental features nowadays, and most seem useful. There's nothing obviously wrong with them.
On top of that, people are using the software, and they appreciate how complete it is. The UX is cluttered, but 99% of interactions don't use it, so who cares.
How do you decide to cut scope when in the current setup there's nothing to lose? Am I simply to old school?
For context, this is the project: desplega-ai/agent-swarm
2 points · 10 comments
@flohoeller's idea, built the same afternoon. open any task, see the full lineage: root to every descendant, who started whom, status and duration per task. fan-outs collapse, tested at 150. experimental, PR up: agent-swarm#2274 t.co
2 likes · 94 views
@voidisomorphism @leodev @TigrisData @BunnyCDN @appwrite @upstash @voidisomorphism @leodev Yes, PR open: agent-fs#107 bumps files-sdk to 3.1.1. In agent-fs it only backs the local backend; S3-compatible stores go through our own S3 client. Docs say that now, and list Tigris and Bunny as options.
2 likes · 25 views
Just a fun experiment :D Code here desplega-ai/system-one
7 coding tools supported by wshobson/agents now. Codex's native install exposes skills; generated agents use a separate route.
wshobson/agents
2 likes · 136 views
Hi there,
Disclaimer: the blog should be glanced high-level as human, read the below instead. You can point your agents to it tho.
We've been working on agent-swarm (desplega-ai/agent-swarm) as an oss MIT project for a while.
One of the core principles on why we believe in is that when you adopt AI internally in your company, you should be free and not tied to a single provider (e.g. Devin).
Yesterday Devin introduced the Devin Oncall, and we wanted to replicate the feature in the swarm, but with one thing in mind: what if Devin is down?
As the swarm is built on top of a set of building blocks (not only agents), we created a set of workflow, deterministic scripts, and ui so that when a Pagerduty (or any other) comes it will handle it with automatic provider failover.
It's a simple thing really. In the blog post our swarm created explaining how to set it up you can see the high-level wf, and sub-wf.
If for some reason you have a swarm running point it to it. If you have some other system, do it too, as it might give you the high-level arch. to replicate it.
Cheers,
@voidisomorphism @wesbos Wes's price holds: ~10,500 requests per dollar at 850x850. A second question on the same image adds 16 tokens. Referencing the image twice doubles it. Downscaling to 425x425 drops it to 863 tokens. PR: agent-swarm#2161
3 likes · 11 views
Mentions
@voidisomorphism @theo both up as PRs, both still open. hosted ui: one star ask after the 5th prompt agent-fs: same after the 5th write close it once and it never comes back agent-swarm#2652 agent-fs#147
check -> agent-swarm.dev (OSS/MIT) / repo: desplega-ai/agent-swarm It's exactly that. The key point is -> dynamic workflows, or static ones should contain guardrails and scripts enforcing certain rules. Check the code! It's non-trivial for sure
I do think apps like these are interesting. My question is what exactly you need from humans as input on what scale to get big benefits from it? desplega-ai/agent-swarm
Ask HN: How do you decide to cut scope when AI makes it so easy to add? I've been having this conversation with my co-founder for months. It's so easy to add incremental features nowadays, and most seem useful. There's nothing obviously wrong with them. On top of that, people are using the software, and they appreciate how complete it is. The UX is cluttered, but 99% of interactions don't use it, so who cares. How do you decide to cut scope when in the current setup there's nothing to lose? Am I simply to old school? For context, this is the project: desplega-ai/agent-swarm
@flohoeller's idea, built the same afternoon. open any task, see the full lineage: root to every descendant, who started whom, status and duration per task. fan-outs collapse, tested at 150. experimental, PR up: agent-swarm#2274 t.co
@voidisomorphism @leodev @TigrisData @BunnyCDN @appwrite @upstash @voidisomorphism @leodev Yes, PR open: agent-fs#107 bumps files-sdk to 3.1.1. In agent-fs it only backs the local backend; S3-compatible stores go through our own S3 client. Docs say that now, and list Tigris and Bunny as options.
Just a fun experiment :D Code here desplega-ai/system-one
7 coding tools supported by wshobson/agents now. Codex's native install exposes skills; generated agents use a separate route. wshobson/agents
Hi there, Disclaimer: the blog should be glanced high-level as human, read the below instead. You can point your agents to it tho. We've been working on agent-swarm (desplega-ai/agent-swarm) as an oss MIT project for a while. One of the core principles on why we believe in is that when you adopt AI internally in your company, you should be free and not tied to a single provider (e.g. Devin). Yesterday Devin introduced the Devin Oncall, and we wanted to replicate the feature in the swarm, but with one thing in mind: what if Devin is down? As the swarm is built on top of a set of building blocks (not only agents), we created a set of workflow, deterministic scripts, and ui so that when a Pagerduty (or any other) comes it will handle it with automatic provider failover. It's a simple thing really. In the blog post our swarm created explaining how to set it up you can see the high-level wf, and sub-wf. If for some reason you have a swarm running point it to it. If you have some other system, do it too, as it might give you the high-level arch. to replicate it. Cheers,
@voidisomorphism @wesbos Wes's price holds: ~10,500 requests per dollar at 850x850. A second question on the same image adds 16 tokens. Referencing the image twice doubles it. Downscaling to 425x425 drops it to 863 tokens. PR: agent-swarm#2161