@Tiny_Fish Yes, awesome. Need access, I have a social impact project in mind
English
Arnab Saha
95 posts
Arnab Saha
@arnandroid01
#marketing #startups
Chennai, India Katılım Aralık 2013
1.1K Takip Edilen50 Takipçiler
We are building something super huge in public!
An open source project that lets you build datasets on literally anything, on-the-fly, with fresh live data on the web.
AND it gets updated periodically.
If this excites you, even a little, DM us and we'll get you onboarded!
Simantak@not_simantak
Introducing: BigSet. A new open source project, that lets you ask for any dataset and watch agents to assemble it for you! Live datasets, on any topic, that update periodically! Repo & Discord ⬇️
English
to celebrate 3 months since lauching @lucent_ai, we're giving away 5 Codex Pro / Claude Max plans 🎁
to enter, like this post + comment which one you'd pick (codex vs claude)
winners will be selected from comments in 5 days 🫶
English
@MilksandMatcha @MistralAI I will make mistral good in agents if selected
English
Giving away 3x $100 @MistralAI Credits
They've released a lot of really cool features recently. Comment why you want it, 3 winners picked in 48 hours.
Sarah Chieng@MilksandMatcha
I have worked at 7 companies since graduating from MIT. Every single one of those jobs I got by sharing my personal website. LinkedIn and resumes have been pretty irrelevant for me. The modern way to get a job is to create a portfolio. A website that > conveys who you are > projects that you've worked on > what you invest your time into communicates infinitely more than a stiff resume or corporate-pilled LinkedIn. I built my personal website from scratch in less than 5 minutes with @MistralAI's Mistral Vibe. Available through the Le Chat Pro and team plan Get started with Mistral Vibe: mistr.al/vibe-sarahwu-x :) #MistralAI #MistralVibe
English
@OfficialLoganK Google billed me $50 per Gemini API call. That's not a pricing tier. That's a bug. Or a leaked key. Or both. ₹1.26M ($15K) in 24h. Service paused. We don't even use the model that drove 60% of the charges.
Can you please help? Case: console.cloud.google.com/support/cases/…
English
@Teknium 🐛 Bug report: hermes update fails (Access denied, hermes.exe in use). ddgs install fails (venv lacks pip). pyproject.toml exclude-newer = "7 days" unparseable by uv 0.9.3. #HermesAgent
log: pastebin.com/1sArR9JT
English
Have a few more Max subscriptions to give out to folks if you missed this!
Just reply and DM me :)
Cognition@cognition
Intelligence at 1000 tokens per second, right in your terminal. Now available with SWE-1.6 Fast, powered by @cerebras. We're giving the first 100 people who respond a free month of Max to try it out.
English
TIME FOR ANOTHER GIVEAWAY
3 people will win 25$ in Deepseek API credits each (About 1 Billion tokens per person over the next month)
I will help each of you setup Deepseek into your harnesses.
Winners selected from comments, priority to:
- OSS contributors
- Funny comments
English
Giving away @CommandCodeAI Max subscription to someone at random who follows me and Command. RT.
That’s more than 5 billion tokens of DeepSeek v4 pro.
In 24hrs. LFG!!
Read the eng deep dives below, good for all not just us.
Ahmad Awais@MrAhmadAwais
interesting milestone: @CommandCodeAI on pace for ~1,000 new subs/day today. the broader thought imo is that devs are figuring out that running open models inside Claude Code was leaving a lot on the table. seeing more and more posts of DeepSeek/Kimi beating Opus/GPT once you swap them into Command Code instead. the harness matters, often as much as the model itself (a point that i think is still pretty underrated). my eng notes on our harness engineering below.
English
Who wants to beta test a $1/mo coding agent Go plan we’re launching at Command Code next week?
English
Arnab Saha retweetledi
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.
🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.
Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today!
📄 Tech Report: huggingface.co/deepseek-ai/De…
🤗 Open Weights: huggingface.co/collections/de…
1/n
English
@MilksandMatcha Hoping to make a difference in Healthcare AI non-HIPAA use case using Codex Pro! Will appreciate the trial
English
Giving away 5 more Codex Pro plans for folks to try out multi-agent workflows with Codex and Codex Spark
Each person will get 3 months of free Codex Pro (highest tier).
Winners will be selected from comments in 48 hours, comment below why you want it.
Sarah Chieng@MilksandMatcha
English
@MilksandMatcha @cerebras I want a good David vs Goliath story, and will do my part 😁
English
Giving away 5 Windsurf Max ($200/month) plans
Each person will get 3 months of free Windsurf Max (highest tier). Try out SWE 1.6, Cognition's latest, fastest, and most intelligent model, powered by @cerebras.
Winners will be selected from comments in 48 hours, comment below why you want it.
Cognition@cognition
We’re releasing SWE-1.6, our best model in both intelligence & model UX. SWE-1.6 matches our Preview model on SWE-Bench Pro while dramatically improving on various behavioral axes. It’s available today in Windsurf in two modes: free tier (200 tok/s) and fast tier (950 tok/s).
English
@MilksandMatcha I have tried the codex free tier and the usage limits are pretty surprisingly generous. I want to try out making the pro tier sweat!
English
Giving away 5 Codex Pro plans
Each person will get 3 months of free Codex Pro (highest tier).
Winners will be selected from comments in 48 hours, comment below why you want it.
OpenAI@OpenAI
Today, we closed our latest funding round with $122 billion in committed capital at an $852B post-money valuation. The fastest way to expand AI’s benefits is to put useful intelligence in people’s hands early and let access compound globally. This funding gives us resources to lead at scale. openai.com/index/accelera…
English
English
English
One of the immediate use cases for optimize-everything is this demo that autogenerates Skills, given a repo: gepa-ai.github.io/gepa/blog/2026…
English
I'm giving away my entire @openclaw architecture. Behind my $250k/month agency.
After weeks of building, I've dialled in the exact system that runs my business 24/7.
What's included:
• Memory folder structure (how to organize agent context)
• Cron job templates (daily briefs, meeting syncs, content automation)
• How to build a custom dashboard in @lovable
• API reference doc (so your agent never forgets its tools)
• Voice training method (85 posts to teach it your style)
• Supabase schema for dashboard connection
Comment "OS" and follow. I'll DM it to you.
P.S. This will probably blow up so give me some time to reply.
English
Stop paying for LoRA training.
ModelScope just dropped a "100% Free Civitai" alternative called Civision. It combines Hugging Face + Civitai features and at zero cost.
→ Train Qwen-Image LoRAs online
→ Free API inference (DeepSeek, Qwen)
→ No GPU costs
English
@DadBodGeoffTV @Vtrivedy10 Okay this is very nice! Would you continue to maintain it?
English
Hey!
I OSS a project called drift that does this for you but incorporates it back to your agent through MCP or CLI
It utilizes AST parsing with a regex hybrid fall back to index codebases into 15 categories with 400 patterns detectors and also utilize call graph analysis.
I’ve also built drift cortex which is an added persistence layer on top that allows you to index and save notes with an adaptive and learning memory layer that utilizes episodic, core and tribal memory 7d, forever and 365d life spans. Also used through CLI and MCP
All of drift core is capable to run fully offline and can fully set up your source of truth with no internet or ai required github.com/dadbodgeoff/dr…
English
for codebase search i’m more bullish on:
1. spend a lot of compute up front to build a good codemap index in markdown (ex: Deep Wiki)
2. Be very thorough in updating this index using git diffs and an intelligent model. This is the source of truth for search.
3. Agents use targeted+parallel grep using the codemap md file as a reference
spent like 2 years grinding on semantic search/retrieval (disclaimer it was for vision), there’s a use case for it and maybe can boost perf for code search
But we can trade the compute in indexing via embeddings for exhaustive search with Agents and the architecture is much simpler for >90% of use cases
Models are great at reading text and following diffs so let them read. And markdown is way more interpretable than embeddings
Ethan Lipnik@EthanLipnik
Does anyone know why Codex and Claude doesn't use cloud-based embeddings like Cursor to quickly search through the codebase?
English
@arnandroid01 hello! please send me your request id so we can have a look
English
Free training for Z-Image is now live on our international site! Open to everyone—go check it out! modelscope.ai/civision/model…
👉Any question? Join our Discord:
discord.com/invite/hgKcfgX…
ModelScope@ModelScope2022
🚀Introducing Z-Image – the full, non-distilled version of Z-Image for high-quality, customizable image generation. ✅ Model: modelscope.cn/models/Tongyi-… ✅ Demo: modelscope.cn/aigc/imageGene… ✨ Capabilities: • Photorealistic + diverse art styles (anime, illustration, digital art) • Full CFG & precise negative prompts for reliable prompt following • Multi-subject generation with distinct identities—no "clone faces" • Community-friendly: ready for LoRA, ControlNet, and other fine-tuning workflows 🔥 Highlights: • Full architecture (vs. distilled Turbo): better diversity & training stability → ideal for customization • High output variance across seeds = less homogenization • Built on Scalable Single-Stream DiT (S3-DiT): more parameter-efficient than dual-stream designs • Trained via SFT: balances visual fidelity and fine-tuning ease 🛠️ Tooling & Ecosystem: ModelScope DiffSynth-Studio➕ Includes Z-Image-i2L: generate a custom style LoRA from just one image ✅ Model: modelscope.cn/models/DiffSyn… ✅ Demo: modelscope.cn/studios/DiffSy…
English