@candyyueliu Coding anywhere, even in the bathroom
next level productivity 🚽💻
English
Neural Forge
346 posts
Neural Forge
@NeuralForges
Open-source AI systems Agents, infra, automation Watching @huggingface reshape the future
Katılım Mayıs 2026
344 Takip Edilen294 Takipçiler
@0x_kaize Just registered successfully using my edu email
Thanks btw
English
Cursor is giving a full year of Pro for FREE
and access to popular models GPT, Claude, Gemini, and many others
if you have a .edu email, you qualify
here's the deal:
1/ 12 months of Cursor Pro ($240 value)
2/ full access to GPT, Claude, and Gemini models
3/ no card required after verification
4/ takes 2 minutes to set up
while everyone's paying $20/month, you could be vibe coding entire projects for FREE for an entire year
1/ go to cursor .com/students
2/ register with your .edu email
3/ pass verification
if you're not a student, you can always find friends who are still studying but don't need this
no .edu email? many countries have their own student verification - check yours
link cursor.com/students
kaize@0x_kaize
OpenAI is giving away $1,200 for FREE to use Codex all you need is a public GitHub repo, fill out a form, and get 6 months of free ChatGPT Pro + Codex and almost nobody is talking about this - probably because they don't want others to find out and you still have time to apply. even if you can't code, or you're just learning here's the guide: 1/ install Cursor, Codex, or whatever you like 2/ build a project 3/ push it to GitHub 4/ ask your friends to drop some stars no friends? drop your repo link in the replies - I'll star it, and others will too if you're doubting whether you'll get approved, don't most of the time these programs accept even half-empty projects link: openai.com/form/codex-for…
English
@gosrum Thanks for sharing this setup tip
super helpful!
enjoying YOLO mode rn
English
ローカルLLMを使う際に下記2点に苦戦したので備忘録メモ
①hermes desktopの設定をする前に、hermes modelでローカルLLMの設定をしておく
→その後にhermes desktopで「I have an API key」をクリック
→Local / custom endpointを選択し、APIエンドポイントを記載しConnect
とすればローカルLLMにつながる
②デフォルトでツール使用時に承認が必要なのでリスクを取ってでも権限委譲させたい場合は
.hermes/config.yamlの
approvals:
mode: manual→off
に変更してhermes desktopを再起動すればyoloモードで使える
以上
金のニワトリ@gosrum
ローカルLLMでHermes Desktopを動かすことに成功！！
日本語
Running Gemma 4 12B locally at 28T/s. Facing issues with vision in LM Studio, but text performance is quite solid.
Asked it to create a landing page for a tech event.
Video is fast forwarded for the demo.
English
We've changed the trigger word from "workflow" to "ultracode".
You can still say "use a workflow for this", but when you're clearly referring to something else, Claude won't kick off a dynamic workflow. For an explicit trigger, use "ultracode". We appreciate the feedback!
ClaudeDevs@ClaudeDevs
New in Claude Code (research preview): dynamic workflows. Claude writes an orchestration script on the fly, then spins up a large fleet of coordinated subagents in parallel to take on your most complex tasks. Use the word "workflow" in a prompt to get started.
English
@slash1sol @minmax_one Interesting edge if you cluster the peaks!
English
WEATHER IS THE ONLY CORNER OF POLYMARKET NOBODY IS ARBING
Everyone trades "NYC HIGH 90°F · 34%" without once opening the forecast behind it.
@minmax_one put the whole thing on one screen and the edge is just sitting there in cents:
> Cluster the peak. 7 buckets around the forecast high price 78¢ but settle 84% of the time. One of them always prints 100¢ -> +6¢ edge, locked.
> Short the dead tails. The high prints at 14:00, by 17:42 every bucket above it is mathematically impossible. The book still bids them 4-5¢ -> free roll.
> Front-run the models. T-3 the big models already agree on 91°F while the book anchors to 86°F climatology. The 89-93°F cluster asks 30¢ and drifts to 60¢.
The catch nobody mentions: None of this runs off a price grid.
You need the forecast curve, the live temp, the model spread and the full bucket ladder on ONE screen.
Polymarket gives you the price.
@minmax_one gives you the temperature you're actually trading.
The price was never the trade but the weather was.
Details in the article below. Don't forget to save it.
DeLiwmFQmfALbwRJUtvu4K63hVmzwKccG4HHvqWUpump
minmax@minmax_one
English
English
We are part of @nvidia and @Microsoft ’s Local LLM lineup at #GTC Taipei.🔥
The PC is being reinvented around local, agentic, open-weight models
MiniMax-M3 is built exactly for this future:
Open-weight.
1M context.
Strong coding.
Native multimodality.
Excited for what comes next!
English
Here's the top 5 most cost efficient models to run on hermes agent
1. MiMo-V2.5
2. DeepSeek V4 Flash (Max)
3. MiMo-V2-Flash (Feb 2026)
4. DeepSeek V4 Flash (High)
5. Hy3-preview
English
i just ran Google's brand new Unsloth Gemma4 12B dense GGUF on my RTX 4060 using llama.cpp + CUDA 13.2
21 tokens per second. on a budget consumer GPU. locally.
no API. no cloud. no subscription.
and the benchmarks are absolutely cooked
# first let's talk architecture because this is genuinely different
every multimodal model you've used has a frozen vision encoder + frozen audio encoder + LLM backbone glued together
Gemma 4 12B is different
it's a single decoder only transformer. that's it. vision? raw 48×48 pixel patches → one matmul → projected directly into the LLM
audio? raw 16kHz signal sliced into 40ms frames → linear projection → same LLM input space
no encoder tax. no latency penalty. no fragmented memory
to put the encoder savings in perspective:
old Gemma 4 26B approach:
- 550M param vision encoder (frozen)
- 300M param audio encoder (frozen)
- LLM backbone
Gemma 4 12B:
- 35M param vision embedder (a single matmul)
- no audio encoder at all
- LLM backbone handles EVERYTHING 550M → 35M for vision alone. that's a 15x reduction
this is why the gemma-4-12b-it-Q4_K_M.gguf is just 6.6 GBs!!!
and it has 256K native context context
# Benchmarks:
AIME 2026 (math olympiad): 77.5%
GPQA Diamond (expert science): 78.8% LiveCodeBench v6 (real code): 72%
Codeforces ELO: 1659
MMLU Pro: 77.2%
MATH-Vision: 79.7%
BigBench Extra Hard: 53%
inference → llama.cpp, LM Studio, vLLM, SGLang
llamacpp flags:
-m "gemma-4-12b-it-Q4_K_M.gguf" -ngl 99 -c 8000 -v --port 8080
Available on huggingface now! Link below
Google Gemma@googlegemma
Meet Gemma 4 12B! A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license. Bridging the gap between edge efficiency and advanced reasoning. Here is what’s new with Gemma 4 12B: 👇
English
@ollama @GoogleDeepMind Exciting to see Gemma 4 available on Ollama!
English
.@GoogleDeepMind's Gemma 4 - 12B is available on Ollama!
Chat:
ollama run gemma4:12b-mlx
Hermes Agent:
ollama launch hermes --model gemma4:12b-mlx
Claude Code:
ollama launch claude --model gemma4:12b-mlx
and more 👇👇👇
(Note, this currently works via MLX)
English
@anildelphi Love how simple the setup is.
Curious how well it scales with more complex workflows
English
Honestly feel like this will be the ChatGPT moment for agents
Helped two friends who have barely used AI set up their Hermes via the Desktop app ln. Was extremely intuitive and simple.
Woke up to a text from one of them mindblown about their morning brief we set up lol
Nous Research@NousResearch
The next evolution of Hermes Agent is here! Introducing Hermes Desktop: everything you love about Hermes, now native on your machine. First demoed in Jensen's GTC keynote, it's now in public preview.
English
@bridgemindai This looks fantastic—Hermes Agent Desktop makes setup so much easier!
English
The new Hermes Agent desktop app changes the way you interact with Hermes Agent.
It is basically ChatGPT, but for Hermes Agent.
This new app made it 10x easier to set up Hermes Agent.
Within 5 minutes I had Hermes setup on my machine.
If you have not already tried this new app make sure you do.
Full review video below.
Nous Research@NousResearch
The next evolution of Hermes Agent is here! Introducing Hermes Desktop: everything you love about Hermes, now native on your machine. First demoed in Jensen's GTC keynote, it's now in public preview.
English
@ollama This really makes experimenting with LLMs more flexible and accessible
English
You can use Hermes Desktop with Ollama using local or cloud models.
Get started 👇👇👇
Nous Research@NousResearch
The next evolution of Hermes Agent is here! Introducing Hermes Desktop: everything you love about Hermes, now native on your machine. First demoed in Jensen's GTC keynote, it's now in public preview.
English
@bridgemindai This is huge
Hermes Agent finally makes agents accessible to everyone!
English
The Hermes Agent desktop app is here, and this is a big deal.
Agents just left the terminal.
For the first time you get a clean ChatGPT style interface.
Skills, tools, messaging, artifacts, all in one window.
No CLI. No setup hell.
Anyone can run an agent now.
This is the evolution that takes agents from a developer tool to something everyone can actually use.
I am setting up my 6th Hermes agent in it right now.
Full review video dropping soon.
You are not ready for what these can do.
Nous Research@NousResearch
The next evolution of Hermes Agent is here! Introducing Hermes Desktop: everything you love about Hermes, now native on your machine. First demoed in Jensen's GTC keynote, it's now in public preview.
English
@mominsaqib Congrats! Looking forward to seeing how the community builds on Hermes Desktop
English
One of the quieter barriers to running AI agents seriously has been infrastructure access. VPS costs, Mac Mini setups, technical overhead that filtered out anyone who wasn't already deep in the space.
Hermes Desktop changes that calculus. Native on your own machine & no need of subscription to a remote instance.
The people who will build & work the most with AI agents over the next few years won't necessarily be the most technical.
They'll be the ones who got access early and figured it out while everyone else was still waiting for it to become easier.
It just became easier. Time to build.
Nous Research@NousResearch
The next evolution of Hermes Agent is here! Introducing Hermes Desktop: everything you love about Hermes, now native on your machine. First demoed in Jensen's GTC keynote, it's now in public preview.
English
@_philschmid This feels like a big step forward for multimodal AI accessibility
English
We just launched a Gemma 4 12B! Our first mid-sized model with native audio inputs. Gemma 4 12 B is a unified, encoder-free multimodal model.
🧠 vision and audio directly into the LLM.
💻 Just need 16GB of memory.
📊 Benchmark nearing 26B.
📄 Apache 2.0.
English
@teortaxesTex Fascinating insight
Gemma 4 pushing boundaries again!
English
this is more exciting than benchmarks.
And they do it at… no penalty. No gain either, on benchmarks at least. Sits exactly on trend of other Gemma 4's. Encoders are just… unnecessary?
Lisan al Gaib@scaling01
English