ThisAndThat

@jxnlco I started using Codex a few days ago and it has been fixing dozens of bugs that were created by Opus 4.6 and Opus 4.7 AND does what I tell it! I am really impressed and can't believe it. Former GPT user driven to Claude by the propaganda.

jason from the codex team here, heres a draft on codex maxxing and the primatives i use on a daily basis jxnl.github.io/blog/writing/2… would love any feedback

@azerX @Teknium The themes are all terrible. A great product this Hermes but the ui is just unusable for me. Well I sent Hermes to create a more readable theme. After 150 rounds it came back with a theme not much better. Thats where I uninstalled it but later gave Hermes a new try.

I love the idea of Kanban but its very very badly designed! I can't use it as it is. the cards and content is (to me) unreadable! There should be better theme and font control. Having all capital and this way of presenting the result...well just look at the image.... Love hermes but kanban to me is the weakest feature with the UI.

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…

@Teknium Damn it. If I knew you would create that so soon I wouldn't have spent days to make my custom addition/orchestrator work.

@yacineMTB By the way I started using GPT 5.5/Codex 2 days ago. It has been fixing dozens of issues that were messed up by Claude in the last weeks/months. I still can't believe it and extremely happy because I was about to throw away > 200k lines of code.

Apparently, if you're a straight white male, Anthropic throttles your Claude Code tokens for equity reasons.. they're running a prediction model on your computer to do that. that's too much. some people have been using slang from the hood to increase their token allowance

@yacineMTB I was wondering whether Claude purposely destroys my codebase and always tries to lead me to a rabbit hole while telling me I was so excellent or am I paranoid? Well, may be not so far fetched.

@Teknium @steipete Hey guys stop fighting You guys are better than that You both have done the world a great service don't ruin it and all the great stuff is still going on nowhere near being done yet.

I’m sorry that you’re this desperate that you will take such an unscientific benchmark instead of any established one. Also qwen local is one of the most random length models there is with all its looping. And we smoke you all on quality benchmarks on every open model. Here’s wildclawbench by internlm, same speed on open models, much better results.

Looks like our focus on performance paid off.

@Michaelzsguo I think Deepseek is $0.43/1M tokens on openroute. So how does that add up? ~100M tokens ? More like > $43 Euros And how about how many times it was interrupted during the 13 hours, how many times it returned garbage or simply nothing? Propaganda!

My Deepseek V4 Pro agent (inside codex) has been pursuing goal for more than 13 hours, burning ~100M tokens, and has only costed me $1.85. Yes you saw it right. Not $185, but 1 dollar 85 cents.

@ClaudeDevs So marvelous. I suggest Anthropic apply for EU membership. They would fit perfectly. So bureaucratic and annoying. And by the way Opus 4.7 is totally overrated. Not nearly as good as the propagandists make you to believe. Definitely not worth the 2-300% prices..

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

@tolibear_ what is the difference to Hermes kanban, and their dashboard?

Codex /goal now has a native Kanban board. Starting a /goal run now fires up a lightweight Kanban board with clickable cards that move as Codex completes tasks. npx goalbuddy Or update with npx goalbuddy update Start the goal-prep skill

@NousResearch @trycua The question is what is the failure rate? You need only 2 wrong clicks and your computer is gone. May be useful in a virtual machine sandbox but not on your real computer.

Computer use with any model Hermes Agent × @trycua

@Jonnyvandel OPERATOR

plz welcome: masscontent operator sh*ts on openclaw sh*ts on claude code sh*ts on higgsfield im pretty positive that this is the most powerful mass marketing agent on earth. just a short list of what he can do: 
- full phone farm integration (~$4 per account) - second brain on u and your competitors - 500+ marketing .skills pre-loaded - call 64 ai content models (50% of the cost) - build full ai vids/slides workflows in bulk (30 at once) - create consistent influencers and map them to accs - run ai theme pages by itself - auto-post everywhere via our phone farm - setup up lead mag auto-dm’s for every post - spin ai inbox agent to manage every dm and follow up - social intelligence scraping on every platform - clone viral posts with your brand + brain - fully edit every video - highly realistic Seedance 2.0 ugc - render mocked iOS app screen recs w/ your real ui - write tweets & post any render media - create saas mini-launch vids with every tweet (real ui) - scrape top ai influencer pages and clone them - use our credits or connect your anthropic key for unlimited - track posts work & edit all workflows to double down - all analytics mapped in his brain wiki - track every click and conversion event - create & post 1,000+ vids/day i’m not sure how i got here, but 2.1 million lines of code later… we are locked in. like, rt + comment “OPERATOR” and i’ll dm it to you. (must follow for dm)

@ModestMitkus Congrats. Thats a very nice achievement. 1000 users in 4 months. Well done!

My first mobile app just hit 991 subscribers! Only 9 away from 1,000! Stop overthinking, just start.

@OpenAI I am a GPT 5.5 user but Gemini multimodal Live models have been able to do that since November 2024 at least. You're late to the party! What about the pricing?

Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents. Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold. Now available in the API alongside streaming models GPT-Realtime-Translate and GPT-Realtime-Whisper — a new set of audio capabilities for the next generation of voice interfaces.

@sudoingX Well my Hermes failed the first test miserably. Told him to change the dashboard skin to a light/cream one. It took > 60 turns but results were crap. By the way using GPT 5.5! In cursor I do that with 1 short prompt. If thats its capabilities then meh...

if you are running local ai or thinking to start, if i could give you one single piece of advice it is this: choose your agentic harness carefully. it matters more than the model. i have lost count of how many people have dm'd me saying their local model is "dumb" or "broken" or "not as good as the cloud one." then they switch from openclaw or some other bloated framework to hermes agent and the same model suddenly works. just clean tool calls and the agent doing the thing it was supposed to do. hermes agent is the best general purpose agent i have used in 2026. drives my single 3090 with qwen 3.6 27b dense q4, drives my dgx spark with nemotron omni q8, and the same harness handles coding, research, video editing, automation, anything you point it at. packed with skills out of the box (browser tools, code, github, jupyter, multimodal, more than i have used yet), full tool calling that holds across long sessions, persistent memory, sub agents. if you tried local ai once or twice and gave up because it felt half baked, the issue might not have been the model. it might have been the harness wrapping it. swap the harness, run the same model again, and watch what changes. hermes agent is the one i recommend to everyone running local. and especially to anyone who almost gave up on it.

@athcanft Great stuff! Didn't know that this is allowed at all. And thanks for showing your "trade secrets" hhh

i made a tool that makes & schedules 100 viral vids for my app in under 15 mins how it works: → auto scrapes & downloads viral shorts → stitches my app CTA onto the shorts → schedules to TT + IG + YT this is the ultimate form of lazymaxxing organic content

@atomicbot_ai @NousResearch Whats the deal with this atomic thing? I spent quite some time to understand before installing it I mean if it is simplicity of installation. Well Hermes takes a few minutes to install, just answering a few questions So? Local? You can use local models with Hermes directly, no?

Hermes Agent by @NousResearch (100k+ ⭐) now inside Atomic Bot: – Free Local models: Qwen, Gemma or – Use your API keys for any provider – Dashboard, terminal, logs and files explorer – Private and Open Source Download MacOS app or run in Cloud👇

@mr_r0b0t @read_jfk_files ,??

@ThisAndDhat @read_jfk_files Those are the things you keep airgapped 😁 Deepseek V4 didn't invent industrial espionage! Tbh our "smart-plugs" are probably exfiltrating as much if not more PII

Be me: Sign up for DeepSeek API because everyone is saying how reasonably priced it is. Load up $50, let’s see. 30 minutes in, here’s where we’re at using V4-Pro

@hiarun02 They just limited Copilot rates, no? Time for me to cancel that damn thing. Haven't been using it much The extension/agent is so bad Nowhere near Cursor which I use most of the time but their rate limits make it unusable with the pro plan 3 days work and its full Back to auto!

Is anyone still using VSCode instead of switching fully to Claude Code, Cursor IDE, or OpenAI Codex? Or am I the only one?

@HeiderKate I really hope the Chines are fast enough to overtake Canada just to see them fuck up people like you. China is a 100% human abuser and has killed 10's of millions of innocent people And a disgusting person like you they are fine?

@read_jfk_files @mr_r0b0t I wonder whether all these China advocates in this thread don't understand or just play dumb? It's not about source code where a model produces tons of it in a minute It's about ideas and architectures which all AI analysts/CEOs say is the top thing Ideas make the difference

@mr_r0b0t yeah, but you just gave all your source code to China. maybe stop and think about why it's so cheap it might as well be free.