Roger

@0xSero I wonder if it’ll run effectively on a 64GB Studio. Will be trying this weekend

New best local model for y'all 16GB-64GB rejoice, the chosen one has arrived. huggingface.co/Qwen/Qwen3.6-2…

@opinion_bits @_MaxBlade Ollama.com $20/mth. No skills required

@_MaxBlade People use opus4.7 because they can use it off the shelves and not k2.6 because it’s a nightmare for nontechnical people to set it up.

Everyone says Hermes / Openclaw is unusable without Opus 4.7 I just ran two Hermes agents head to head, one with k2.6 and one with opus 4.7, through three heavily agentic tasks with one prompt. OPUS DID NOT COMPLETE 😳 Kimi completed all tasks. It took 4x longer, but is 10x less expensive to run. This is SHOCKING.

@strzibnyj They are getting rid of communities?

X getting rid of communities is a friendly reminder to never build your main thing on someone else's platform.

Do what Klein did for Oil Sands and if there is an opportunity money will follow.

@CodeWithZeee @TheGeorgePu Actually I downgraded to Pro for a few months when I was focused on other things that didn’t require opus, nor a ton of Sonnet.

@TheGeorgePu I’d imagine it’s a very small number of pro using Claude code

Anthropic just pulled Claude Code from the Pro plan. Pro users wanting it need Max now. $100/month minimum. 5x jump. I'm on Max 20x so I'm fine. Flagging for anyone on Pro who's about to find out. No announcement. Just a pricing page edit.

@ollama @NVIDIAAI Which model is the most gpu intensive per use that you host. Just curious!

Scaling up Kimi K2.6 with more NVIDIA Blackwell GPUs! Will be adding more in the coming days. Try it with OpenClaw: ollama launch openclaw --model kimi-k2.6:cloud Try it with Hermes Agent: ollama launch hermes --model kimi-k2.6:cloud Try it with Claude Code: ollama launch claude --model kimi-k2.6:cloud Chat with the model: ollama run kimi-k2.6:cloud

@mojito_flow @venturetwins I agree with you Verdis, children don’t need AI.

@venturetwins I’m very pro-AI but this parent is right. Children should not be exposed to LLMs while their brains are maturing. This will be worse than LLM psychosis. I’m glad I didn’t have access to LLM’s growing up.

I am so sad for this kid

Question: why doesn’t Hermes have a heartbeat? One use case is an email inbox check with a memory of last time.

Tip of the day: Learn about Auxiliary Models and save big money with Hermes Agent! youtube.com/watch?v=NoF-Ya…

@oliverhenry I like it. I'm often someone who chooses non-mainstream choices, so while I'm not a psyop I rarely represent the majority. You do you ;)

Hermes is a psyop, i have not seen a single account recommend it that doesn't look like a bot OpenClaw 4 lyf

Wholly molly is @ollama so slow! Like it’s like watching paint dry! Guys charge more, so the service can be maintained. I’d rather $40-$50 a month and 4x that token rate! Heck I’d even be down to lose to token allotment. Some of us have little time to goof around when we are working and we need to clang and bang. Come out with a price that’s got velocity folks in the bullseye 🎯

@Teknium Definitely confused then. I’m confused how Skills is a loop. Thanks for replying @Teknium; appreciate it.

@roger I think you're confusing memory with the self improvement loop. The self improvement loop is exclusively built around Skills - not the session db search or memory

I have found the “self learning” Hermes everyone keeps talking about is only there if you don’t use the default SQLite fact_store and that you find a way to manage the different.md files which seem to constantly get cluttered with random garbage. NOW important note, I use GLM 5.1 as my daily driver, so if you are hitting Opus instead well then YMMV. Tomorrow I’m going to try get more nuanced on how I utilize my LLM models because I’m not sure how to setup Opus for say a planning session and then maybe GLM for the actions later. This is easy on Claude Code but not (to me at least) straightforward at all if even possible. Also I’m using hindsight now so I’m not sure if the “learning” part is impacted more by the main LLM for the Agent or whether the memory LLM matters most for quality storage. None of that is clear to me. No one is posting hey use this like this or this for best results; like a gist config shared in the community would be great.

@Teknium @yao_ej24569 @WesEarnest I’m going to test this shortly because I’m having the issue you are both talking about. Status still isn’t always clear and what did I say 10-15min ago. @Teknium when you hammer with that many agents, what’s your go to model?

@yao_ej24569 @WesEarnest Have you tried v2 TUI, finalizing next week? if your hermes is updated run `hermes --tui`

One thing I don't like about the Hermes Agent is how difficult it is to scroll up and look at a previous message in the TUI. Using a messaging platform fixes this, but there are times when it is generally faster to use the terminal. @Teknium

@jlongster With this SOC compliance will become a thing. Is that a priority for future implementation?

Excited to release workspace support (beta) in OpenCode 1.4.11: lets you run sessions in git worktrees and even remote environments. * Sessions stay coordinated in one place * Remote OC server keeps them running * Local sync keeps your data with you * Plugins can add any workspace type

@MichaelToddLamb @tonysimons_ Check out tunnelTo.dev very cheap but does the high value stuff that most need. Been using them for 2-3 years I think

already paying for grok, and claude, I am a tech enthusiast, i have a back ground in comp sci though its from 20+ years ago. now I am in the construction industry and seeing claude's capabilities, its only a matter of time before even my 'white collar' job is 100% done by ai ... i don't want to wake up to a thousand dollar overnight api bill lol, i realize there are monthly as you have said, i thought for 'tinkering' and 'learning' i could get it at least setup for 100% local, but WSL seems too bogged or my hardware sucks, i wanted to tinker and learn before investing in a newer and more capably piece of hardware (i'm overdue for a machine upgrade anyways but I dont get on it enough to really warrant it so I was going to get a dedicated ai machine) the 'BUY a GPU' guy is directionally right...

Dear Algorithm, Only show this to: > Hermes Agent builders > Claude Code junkies > Codex fanatics > Founders building in public If you’re building something tonight while the rest of the world is at the bar, let’s be besties. 🫡

@Brian29335497 @BertaProudDad @JackDan110 He also said he would invoke something else to get major infrastructure built; his words as they leave his mouth aren’t worth a dime. Everything he says is a lie; all we can do is wait and see what he actually does.

@BertaProudDad @JackDan110 Carney said he would invoke the emergency act to stop Alberta separation at a stellantis plant in Ontario It’s a direct quote, so yeah, I think he’ll try it

DMs flooded with Carney’s “BOMBSHELL” — he can’t crush Alberta independence with the Emergencies Act. Courts ruled Trudeau’s Freedom Convoy use unconstitutional. Fear mongering has no place. Focus on signatures! Alberta first. #AlbertaIndependence #AlbertaFirst #FreeAlberta #AlbertaStrong

Why I switched to Hermes jury is still out, but so far so good. Still it’s all new stuff, so some early adopter patience is required.

Hermes Desktop v0.5.0 – native Mac app for Hermes Agent. SSH only. multi-profile + full cron manager, now live ✅ NO gateway, NO daemon, remote is source of truth. The official dashboard just shipped too – great for browser admin! @Teknium @NousResearch

@Agent_Crafter @Teknium Interested

Trying to write a tutorial for hermes, anyone interested?