@Teknium I love this idea. How obvious and easy is it to delegate roles to the orchestrators and to the doers?
English
Orthorexic Apopheniac
2.8K posts
Orthorexic Apopheniac
@LittleNonsuch
Only hard money can save humanity.
Katılım Mart 2008
230 Takip Edilen81 Takipçiler
The Hermes Agent Kanban just got a big automation upgrade.
Drop one prompt into the triage, and the orchestrator agent can take it from there - decomposing it into all the subtasks necessary and automatically assigning agent profiles that fit the specialization needed.
You can also now add descriptions for each agent profile, to better assist the orchestrator in determining what tasks should go to which profile!
Docs: hermes-agent.nousresearch.com/docs/user-guid…
PR: github.com/NousResearch/h…
English
@gb_rango This bro is just an engagement whore, he will follow the path of maximum engagement
English
Bryan Johnson will eventually, through rigorous exploration and testing, arrive at the correct experimental conclusion that Christ is the only path to defeating death
Bryan Johnson@bryan_johnson
I've been praying the past few weeks. Unsure why. There's good evidence behind prayer. It mimics breathwork, calming the nervous system, dropping cortisol, and quieting the brain. Daily prayers show lower depression, anxiety, and pain. I'd like to develop a prayer practice. Growing up, the protocol was written for me. Explaining whom to pray to, the structure of the prayer, and the boundary conditions. I don't really know how to pray now.
English
@ivanfioravanti @hnshah Say that again Ivan please. Should we run DS4 for agentic work (like Hermes) at 0.6 in your opinion?
English
Running a LOT of evals on ds4 engine right now.
I'm getting great results from both performance and evals point of you on the M5 Tensor Kernel 🚀
But as you can see here best results is with Tensor Kernel temp 0.6.
I want to thank @hnshah for the two runs on its M3 Ultra 256GB/60 GPU cores.
Now running last 2 runs (temp 0 and temp 0.6) with the PR 185 by @fitchmultz that is correcting some false negatives 💪🏻
The M3 Ultra run at temp 1 below was using it. I hope in a 70/92 for M5 🚀
I will then share the full Claude artifacts on the run comparisons and then I'll start testing the q4-imatrix on M3 Ultra 512GB to compress results.
Running DeepSeek V4 Flash locally, with this intelligence is just AMAZING! Thanks for the amazing work @antirez you deserve all the love and hardware gifted from the community here 🙏🏻
English
English
@LittleNonsuch That is already the case are you sure your on latest hermes?
English
So much good content around Hermes!
Shann³@shannholmberg
how to set up your Hermes Agent control room send this image + the repo below to your agent and it will configure itself based on the blueprint github.com/shannhk/hermes… this is the same architecture I use to run specialist agents across both my agencies
English
Yes, the Acer Veriton GN100 can be serviced through Acer's authorized service centers and official support programs as a professional-grade AI workstation.
It's designed for enterprise use with warranty coverage. User disassembly is limited (RAM is soldered; storage access is debated in reviews—some note removable bottom panel + M.2 slot, others call it fully sealed with zero upgrades). For repairs, contact Acer support with your serial number.
English
@bryan_johnson “Dear Lord, please remove all the microplastic from my balls”
English
I've been praying the past few weeks. Unsure why.
There's good evidence behind prayer. It mimics breathwork, calming the nervous system, dropping cortisol, and quieting the brain. Daily prayers show lower depression, anxiety, and pain.
I'd like to develop a prayer practice. Growing up, the protocol was written for me. Explaining whom to pray to, the structure of the prayer, and the boundary conditions.
I don't really know how to pray now.
English
English
English
@ivanfioravanti Yeap. I switched to DS4. Not gonna switch back. It’s amazing.
English
Every day passing, I'm more convinced that we need vertical local inference engine like ds4 to really ensure focus on a specific architecture and being able to test and optimize it to the max.
There is already complexity and too many variables in play with just one model. 🤷🏻♂️
English
I have two NVIDIA DGX Sparks stacked in my office.
They've been sitting there for a month.
Here's my honest take.
Open source AI is never going to compare to frontier models.
Running quantized Kimi K2.6 and GLM 5.1 locally is cool.
But practical? No. Not even close.
I run all my Hermes agents on GPT 5.5 through my ChatGPT Pro subscription. Practically free.
GPT 5.5 is the intelligent model in the world.
Why would I route serious tasks to a watered down local model?
If you need fast and accurate, you're not using local inference.
You're using GPT 5.5 or Claude Opus 4.7.
I'm not saying this to rage bait.
I genuinely want to know.
Why would anyone serious about vibe coding and AI agents use a local model when frontier is this far ahead?
English
@TheoVon Rogan is such a disloyal backstabbing piece of shit for saying this publicly. All because Theo had the courage to speak up.
English
Joe Rogan just took time out of his podcast to express genuine concern for his friend Theo Von.
He admitted that some of Von’s recent behavior after getting on SSRIs “freaks me out” — especially his comments about suicide.
ROGAN: “Theo Von’s going through the exact same thing and last time he was on the podcast he was explaining it to me.”
“It freaks me out because I know Theo has had conversations before...like even publicly.”
“He had a Netflix taping and it didn’t go well. It was like they actually never...they shelved it. They never used it.”
“And you know there was all these stories from people that were there saying he bombed. I think he just had a kind of a breakdown.”
“And when he was talking to the crowd and there’s a video of it, he said, you know, the people were saying, hey, we still love you.”
“He goes, thank you. Look, I’m just I’m trying not to take my own life.”
“And like you hear stuff like that and you just go like, oh, Jesus Christ.”
“I’ve known too many people that I didn’t think were going to kill themselves and then did.”
“And then he goes down these spirals where he starts talking about world events and freaking out. I’m like, oh, Jesus Christ! Like, I got to help this dude.”
English
Orthorexic Apopheniac retweetledi
The worst part of Covid was that 20 million people died
Lisa Boothe@LisaMarieBoothe
One of the worst parts of COVID was watching Americans surrender their God-given freedoms to bureaucrats who were wrong about everything.
English
Orthorexic Apopheniac retweetledi
Absolute masterclass by Owen Benjamin. He completely shatters the establishment narrative by exposing why the media elite hate conspiracy theorists.
Tucker Carlson is left stunned as Benjamin brilliantly points out that those who hate them are the actual conspirators.
English
@deanmckee757 @Dan_Jeffries1 Feels to me like he is just shitting on LLM’s because of some weird taste in his mouth about their brutal empiricism.
Sounds like he wants some closed form equation he can differentiate or something, I don’t know ¯\_(ツ)_/¯
English
@deanmckee757 @Dan_Jeffries1 Yeah, thanks for engaging, I definitely don’t understand La Cun then. If he has some specific argument for a “better latent space” wherein which reasoning can occur, then he’s gonna have to build it to convince me. I don’t see why LLM can’t / don’t already do exactly this.
English
It's actually more than world models that AI is missing. It's five things:
1) Continual learning, aka the ability to update/change its weights to learn from experience and capture that experience permanently.
An MD file or DB does not count because it might not read it on the next turn and context windows are limited.
2) Long term memory. If it can't continually search relevant experience then it is useless and will make the same mistakes over and over.
3) Ability to reason in embedded space.
4) Non forward pass only architecture. In other words the ability to backtrack to previous tokens and "change its mind as it goes." Right now when it says "wait" it is an illusion in its thought stream. It is not actually returning to something it wrote earlier in its thought sequence.
5) World model. This leads to this. That leads to that. If I do this then this happens.
The various issues I've seen coding with AI show this super clearly.
Sorry I dropped the database.
I shouldn't have reverted all your git commits.
These are mistakes even a junior coder wouldn't make. Even as these little wonderful machines get superhuman, and they are in many ways, they're still super idiotic on another level.
Still wonderful to have them but something is missing from their understanding.
Haider.@haider1
Yann LeCun says you cannot build a reliable agentic system without a world model LLMs don't have world models. They can't predict the consequences of their actions before taking them "they just act, and whatever happens next is someone else's problem" Without that, it's not intelligence
English