[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: OpenClaw.png (53 KB, 960x198)
53 KB PNG
Are you still using it? Do you even remember this thing?
>>
>pic
flash game publisher type logo
>>
>>109475880
yes, unless you setup agent automation yourself openclaw is still very much necessary. it makes any model 1000x more useful. though it's slow as hell.
>>
>>109475880
This is a dead meme. Stop trying to revive it. We are all on the next AI fad now.
>>
>>109476024
what you got going on that needs automation?
genuinely curious.
>>
File: totally normal vcs.jpg (108 KB, 1200x675)
108 KB JPG
>>109475880
>>
>>109476101
it's not automation, it's making the model smarter. you you put in a prompt, a model will first try to expand it into smaller items. for instance if you ask it to implement it will break it down into smaller tasks.

then it will set up a number of models (by making multiple queries) to work concurrently on your task. then it will spawn other models to review what each of the models have done, etc.

usually it depends on what you want. for an implementation, it will use multiple rounds of review to settle on an api. once the api is stable, it will ask models to implement according to the api, then ask other models to write extensive tests. after that it can test that each implementation against tests and start over another round if unsatisfied.

it's also pretty much impossible to pass linters without openclaw or some other kind of multiple rounds of asking the models to fix their code for given linter errors.
>>
>>109476168
Humiliation ritual
>>
File: 1778357323067.png (917 KB, 1384x1244)
917 KB PNG
>>109476270
Imagine fucking believing this bullshit, it's an efficient harness that eats tokens and excels at nothing. This tool is exclusively designed for retards, mid-wits prefer Hermes.

Non retards just use OpenCode or Pi, why would they need bloatware with pre-installed shit they won't use.
>>
>>109475880
Never touched this pozzed shit.
>>
i used it for about an hour then just made my own version with pi instead
>>
File: 1757226980810.jpg (119 KB, 1080x1100)
119 KB JPG
>>109477070
Based
>>109477068
This is correct

Honestly I don't know why and who is shilling this shit because this was always a pozzed retard project
>>
>>109475880
It stops working after a while.
Herms seems to be working.
>>
>>109475880
It was a chinese psyops
>>
>>109477199
it's funny that if this was the actual movie dialogue, people at time wouldn't have gotten it.
>>
open claw is unc shit lil bro
>>
When I get bored and want to be insulted into “just use Claude or Codex bro”. Yeah.
But I keep going back because it’s still different.
>>
Bloat with way too many extensions and plugins, which is a big warning sign. Utter garbage on all levels.
>>
>>109475880
>Are you still using it?
No.

>>109475880
>Do you even remember this thing?
Yes. This is a good example of how the lack of human review of pull requests and reliance solely on AI reviews killed this project. It turned it into an unmaintainable piece of crap. 8 months into its development, it's now almost unused. Sad but true.

By the way, I still don't understand why OpenAI bought this project and hired its author to join their team. Objectively, one of the company's worst investments.
>>
>>109475880
Hermes killed it
>>
>>109481890
what's the point of hermes? i have 20$ suscription for chatgpt Plus so I have access to codex. should I try hermes/opencode/pi? Are they different from codex agent?
>>
>>109475880
No, I have anythingLM but only use it for RAG and TTS
>>
>>109477024
BTW that jeet has gone back to the courts to try and stop his deportation order.
>>
is there a single good vibe coded application that isnt spaghetti? I only remember that ffmpeg gui from a anon but thats vala which is very hard to fuck up.
>>
>>109481910
Hermes let's you code while driving the car
>>
>>109475880
it always fucked itself up

one day scheduled things work, few days later it stops working
>>
>>109475880
Turns out i dont need an assistant because I have absolutely nothing going on in my life chat
>>
>>109481910
You're on fucking /g/ and you're asking, "what's the point of https://github.com/NousResearch/hermes-agent , I can just use Codex lol"
(you can plug your oAI key into hermes or use it to run Codex through the CLI btw)
>>
>>109475880
never tried it
i did see 'hermes-agent' in ollama's CLI options and was curious what it is so i installed it. ended up running it on my home server in docker as my general-purpose chatbot-with-memory

doesnt do anything i cant just go to chatgpt.com to do but its more convenient, acessible via telegram,internet web ui so i can use it anywhere and i put shit like documents/books in it's folder to reference
>>
use case? if it can’t make me money, what is the point?

i realised any of the stuff the meme assistant tools can help with is stuff i don’t want to be helped with nor do
>>
>>109488086
>use it to run _____ through the CLI btw
I ain’t falling for that shit again (I will)
>>
>>109488086
>this pic
https://github.com/NousResearch/hermes-agent/blob/main/sqlite_leak_fix.png

this is fucking garbage and you're retarded for shilling it. holy fuck.
>>
>>109475880
we all moved to hermes
>>
>>109475880
I bought 5 mac minis for this shit



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.