Nvrfound
13.9K posts

Nvrfound
@IEatCodeDaily
To fathom the unfathomable. To seek communion with the Divine. 🇺🇦🕊️🍉🇹🇼🌈🌹⚖️💡⚛️
参加日 Haziran 2019
383 フォロー中174 フォロワー
固定されたツイート

@bagholdermy @ism_sol @NousResearch true to some degree. tp harus konsisten model yg dipakeny.
klw ganti2 suka mabok
Indonesia

@ism_sol @NousResearch Is claim of self learning true? A bit skeptical but wow if true
English

gua 0 skill coding belajar banyak hal baru model Grok 4.20 doang karna unlimited nanya udah cenblu di X
ternyata hermes agent buatan @NousResearch ini bisa ambil model di openrouter, artinya kita bisa pake model free semau kita
gua lebih pilih pake hermes karna self improving skillnya

ISM ☀️@ism_sol
lagi install hermes agent, katanya ini better dibanding openclaw
Indonesia

people ask me this all the time so let me say it out loud. the best model to pair with hermes agent is claude opus 4.7, by far, in order of magnitude. the rest is not even close.
pair opus 4.7 with hermes and you feel the unlock immediately. the self improving loop, the per model parsers, the tool call handling, every part of hermes agent amplifies when the model on the other side is that capable.
hermes agent is one of the most well designed and rapidly contributed open source projects in the space right now. the combination with opus is the current ceiling for agentic work.
SS@sageof6apps
@sudoingX Yea but which model is the best complement for Hermes?
English

@Nandapagi ini boros token gk sih?
sy per pekan habis ~50M token utk ngereview sm generate docs
Indonesia

@IEatCodeDaily Ya bikin abstraction aja biar review nya atomic, pake hard rules yg dienforce pake tool yg logicnya udah disunat
Make states transparent kalo kata abang2 FP mah
Indonesia

@Nandapagi gw yg capek ngereview code mereka 😂
emg perlu bikin company wide standard practice and toolset sih
Indonesia

@IEatCodeDaily Abstraction man, just abstract those tools and choke the AI using that
Kita harus sadis kyk jigsaw

English

No timestamp?
is this pre-lobotomized Opus 4.6 or after it got the snip snip?
Claude@claudeai
Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.
English

@IEatCodeDaily Mirip2 sm gitnexus, tapi bedanya dia ngga building anything at start, murni parsing aja dari treesitter jadinya lebih ringan dan bisa langsung masuk
Goalnya cuman ngarahin AI untuk tau where to look aja dengan memanfaatkan static linking
Indonesia


@neural_avb @0xSero unfortunately nothing beats claude on frontend.
glm is kinda better than gpt but still worse than kimi.
minimax and glm is about the same
English

@elliotarledge but with the context length of an ADHD college student
English

@emilstridell @0xShayan @Teknium yep, and not including the one i spent in claude, opencode go, codex, and gemini
English

@maxmilneaus @0xShayan @Teknium yes, but the mcp is kinda crap.
having different models doing vision also often result in shitty result
English

@0xShayan @IEatCodeDaily @Teknium Hermes supports dedicated auxiliary models for eight task types:
1. vision
2. web_extract
3. compression
4. session_search
5. approval
6. skills_hub
7. mcp
8. flush_memories
You can nominate one for each.
And doesn't the Plans come up MCP support for images?
English













