Best Uncensored AI models now :
Overall : Qwen3.8-27B Abliterated
Small GPU : Gemma 4 12B Abliterated
Fastest : Gemma 4 26B-A4B Abliterated
Agentic : Qwen3.6-35B-A3B Uncensored
Writing : Mistral Small 3.2 Abliterated
No surgery : Hermes 4.3-36B
Image : FLUX.1-dev
Video : Wan 2.2
Ceiling : GLM-5.1 Abliterated (754B, not for home)
Where to get them :
all LLMs → huggingface.co/huihui-ai (search the name, grab the GGUF)
Hermes → huggingface.co/NousResearch
FLUX / Wan → ComfyUI + the model's HF page
Setup :
ollama pull hf.co/<repo>:Q4_K_M
ollama run <model>
/set nothink
that's it.
no API key, no account, no card, just few clicks setup
more on how to setup below:
Here's 4 steps to run an uncensored AI on your own GPU for $0:
I run a local abliterated Gemma 4 for anything I don't want leaving my machine
By the end, you'll know how to:
- run a model which has no refuses
- plug it into Claude Code
- do it all for $0, no API key, no account
So, let's go step by step:
1. Install Ollama
curl -fsSL ollama.com/install.sh | sh
you want a recent version, older builds don't speak the Anthropic API and don't support Gemma 4
2. Pull the model
ollama pull igorls/gemma-4-12B-it-heretic-GGUF
the biggest mistake people make here is grabbing a random "uncensored" tag with no quality numbers
this one is abliterated with Heretic. refusals went 99/100 → 0/100 at a KL divergence of 0.028 from the original, so you keep almost all of the base model's capability
one thing: it's still Google's Gemma license underneath, fine for personal use, but read the terms before you build a product on it
3. Turn off thinking (the part every other guide skips)
Gemma 4 is a hybrid thinking model and the abliteration targets non-thinking mode
leave thinking on and the reasoning block eats your token budget and the answer gets cut off
ollama run igorls/gemma-4-12B-it-heretic-GGUF
/set nothink
if you're calling the API instead, add "think": false to the request
4. Plug it into Claude Code (OPTIONAL)
ollama launch claude --model igorls/gemma-4-12B-it-heretic-GGUF
or manually:
export ANTHROPIC_BASE_URL=http://localhost:11434
export ANTHROPIC_AUTH_TOKEN=ollama
claude --model igorls/gemma-4-12B-it-heretic-GGUF
two gotchas:
- bump the context first. Ollama defaults to 4k and Claude Code's system prompt alone is bigger than that, so the model never sees its tool instructions. restart the server with OLLAMA_CONTEXT_LENGTH=32768 ollama serve (setting it in your shell does nothing if Ollama is already running as a service). this needs more VRAM than the 8GB below
- Gemma 4 tool calling is still rough, you should expect some failed edits, imo for serious agentic work use qwen3.5 or glm-4.7-flash instead
What you need:
- 8-12GB VRAM for Q4 (12B)
- ~8GB of disk for Q4, ~13GB for Q8
- more of both if you want a big context
honest take:
- "abliterated" means refusals removed, not quality added. it's still a 12B
- fewer refusals ≠ smarter. for real coding work I still use cloud models
- you own the weights, your prompts never leave the box. that's the actual win
*IMPORTANT: don't use it for anything illegal or harmful. that's on you, not the model
now i'm collecting a list of every uncensored model i'm adding to my local stack and which tasks each one replaces for me
also planning a big "Uncensored 101" article, still figuring out all the angles
save it so you don't lose it
subhojit banerjee☦️ retweeted
i reverse engineered wispr flow, then used deepseek to fine-tune LFM3-1.7B. the result is on par with the wispr flow models you pay for, and it's free, open source and runs entirely on your mac.
key-up latency is 65 ms on my m4 max vs ~520 ms for wispr flow on the same clips. a 6 minute yap session pacing around my office still lands in about a quarter second.
> rust backend with a native metal engine: each decode step runs as one fused megakernel
> a dflash2 speculative drafter is built into the engine, trained on my local 4 x rtx pro 6000 server
> parakeet transcribes, LFM3-1.7B refines. both stream while you talk, so key-up only waits on the last chunk
> never leaves your macbook, so slow internet or bad ping is no longer an issue
> import your own dictionary of terms for your workflow
> i picked every default myself. it feels amazing out of the box
> adapts to the app you're in: discord is lowercase casual, slack and email are formal, claude code gets clear prompts
> training data: 77k transcripts of public audio and synthetic dictations, each rewritten by deepseek v4.1 flash 1017 following a written spec of what "what you meant" means, in 4 writing modes. none of my own dictations went into it
> blind judge (claude opus) on an hour of lectures: phonon's text preferred 90 times, wispr's 23
> built with claude opus 5.5 [xhigh]
brew install --cask infatoshi/tap/phonon
subhojit banerjee☦️ retweeted
A must-bookmark for founders who want to build apps with zero coding experience.
@LinoLeighton’s exact playbook on how he built a $50K/month abs-scanning app at 18 with no business or coding experience.
💸 Update: My total biz expenses have now dropped to just ~$300/mo!
This is to run all my websites/apps serving per year:
- 20,000 paying customers
- 252 million visitors
- 2.4 billion requests
- 2.4 petabyte (PB) in bandwidth
(~$8,000/mo if you add the $7,000/mo GPU bill and $800/mo R2 storage I pay for Photo AI, I exclude it above because it's one project that's so much more expensive in size than all the other sites together, but included in the chart below)
Claude Code helped me to find smart ways to save more money, AI is just really really really good at finding cost savings by itself, which is not so great for the economy short-term, but long term it trims all the fat and opens up spending for better investments I think
Anyway here's what it saved me in money, almost $34,000/mo:
- $29,000/mo - switched to better and cheaper video models on Photo AI, video was very expensive and since I didn't update it ever it stayed expensive, also routed some image jobs to Google Cloud where I have some Google Cloud Credits, and also routed some jobs to Sume which sells AI models at market cost, very cool! This made my GPU bill go from $36,000/mo to $7,000/mo
- $3,000/mo - consolidated my servers, cancelled Digital Ocean and Linode and moved everything to Hetzner saving about $3,000/mo
- $1100/mo - analyzed my Backblaze B2 server backups, I had infinite versioning and lifecycle policies, that's reduced now to 30 days, this literally pushed the bill from $1,200/mo to $26/mo, about 50x cheaper! I had essentially versions of files for all my servers for over a decade, crazy!
- $400/mo - similarly added Lifecycle Policies to my Cloudflare R2 bucket for Photo AI, pics are automatically put in cheaper Infrequent Access after 90 days, there was years of photos in there so this reduced the R2 bill from $1,200/mo to $800/mo
- $118/mo - Claude analyzed all my Cloudflare Domain registrations, lots of domains I didn't use and didn't make sense to keep paying for, lots of things that also don't really make sense anymore in a post-AGI world, renewals for .io TLD are now $50/year which adds up fast!
- $100/mo - Claude also analyzed my Claude subscription (yes!) and discovered I don't even use the Max 20x plan limits, so I downgraded from $200/mo to $100/mo
Even including Photo AI, I have now cut my costs by 90% from last year, getting me close to my biz expenses amount from a decade ago when I started building things
Costs slowly creep up over the years and you don't realize all the things you're spending on but AI is nice to help you figure all that out (also in your personal life!)
✨ I replaced all these SaaS with my own vibe coded now, so about $25,000/mo savings:
- Weather API -> My own vibe coded weather service using Norway's meteorology API
- Image resizing API -> My own vibe coded image resizing API with NodeJS's Sharp + Redis (saves ~$1,500/mo)
- Google Vision NSFW detection -> My own $0/mo vibe coded NSFW detection service with Python's NudeNET (saves ~$2,500/mo)
- Cloudflare Streaming -> My own $0/mo Nginx VPS video streaming (saves ~$250/mo)
- API Flash -> My own $0/mo screenshot service (for social images) (saves ~$250/mo)
- New Relic + IPHub -> My own $0/mo error monitoring, DDOS blocking and fleet management service (called error2telegram) (saves ~$250/mo)
- Human moderators -> My own moderation on all my sites with @xAI (saves ~$10,000/mo)
- IFTTT -> Telegram bot (saves ~$100/mo)
- Photopea as editor inside Photo AI -> Vibe coded my own photo editor (saves ~$100/mo)
- Mapbox -> Vibecoded my own free replacement with OpenFreeMaps (saves ~$500/mo)
- Uptimerobot -> My own $5/mo Uptime Canary app running on Kuma (saves ~$50/mo)
- Ghost blog -> My own $0/mo blog platform
- Customer support -> @featurebasehq + AI support bot (saves $5,000/mo)
- Scrapingbee/SerpAPI -> My own $1/mo scraper with Patchright (save ~$250/mo)
I think I have reached my final limit of replacing my SaaS services as I have these left, but that'd mean becoming an ICANN registrar (I don't want 😁), managing my own email (I don't want), managing my own hard drives at home in a storage (I don't want) and managing my own servers at home too (I don't want) and running my own GPU rig at home (I don't want):
- Cloudflare Domains
- Cloudflare Email Sending
- Cloudflare R2 Storage
- Backblaze B2 for backups
- Hetzner for all my servers for all my sites (VPS)
- xAI for all LLMs for all my sites (so many things like moderation, writing emails, anything)
subhojit banerjee☦️ retweeted
Want to become an AI Engineer in 2026?
Stop collecting AI certificates. Start learning how to build agents that actually DO things.
Bookmark these 6 resources:
1. Hugging Face - AI Agents Course: huggingface.co/learn/agents-…
2. Anthropic - Building Effective Agents: anthropic.com/research/build…
3. OpenAI - Building Agents: developers.openai.com/tracks…
4. LangGraph - Learn Agent Workflows: docs.langchain.com/oss/pytho…
5. Hugging Face - Context Engineering Course: huggingface.co/learn/context…
6. OpenAI - Agents SDK: openai.github.io/openai-agen…
Don't just learn prompting.
Learn tool calling, agent orchestration, memory, RAG, MCP, evaluation, context engineering and deployment.
The goal isn't to build another chatbot.
It's to build AI systems that can research, reason, use tools and execute tasks autonomously.
Bookmark this. Repost it for someone trying to break into AI engineering.
subhojit banerjee☦️ retweeted
Stanford CS329Z: Engineering AI Agents is so underrated on YouTube. It literally walks you through everything you need to build your own AI agent, step by step, hands-on.
Skip Netflix and those endless short videos. Pick a topic that interests you and dive in this weekend. It might just help you land your next $500K job.
youtube.com/watch?v=b4y3rjGp…
course website: cs329z.stanford.edu/
subhojit banerjee☦️ retweeted
This is literally unbelievable…
An 80-year-old with severe Alzheimer’s just REGAINED her memory, emotions, ability to speak, ability to walk, AND bladder control after ONE psilocybin dose.
No Alzheimer’s drug has EVER done anything remotely close to this.
subhojit banerjee☦️ retweeted
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org/
- Excellent recs without any logins (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.
subhojit banerjee☦️ retweeted
Top 15 MCP servers for Hermes Agent, ranked by GitHub stars.
Docs, browsers, code intel, 3D, design + more:
1. Context7
github.com/upstash/context7
2. Chrome DevTools MCP
github.com/ChromeDevTools/ch…
3. Codebase Memory MCP
github.com/DeusData/codebase…
4. Playwright MCP
github.com/microsoft/playwri…
5. GitHub MCP Server
github.com/github/github-mcp…
6. Context Mode
github.com/mksglu/context-mo…
7. Blender MCP
github.com/ahujasid/mcp-for-…
8. Code Review Graph
github.com/tirth8205/code-re…
9. Serena
github.com/oraios/serena
10. n8n-MCP
github.com/czlonkowski/n8n-m…
11. MCP Toolbox for Databases
github.com/googleapis/mcp-to…
12. Xiaohongshu MCP
github.com/xpzouying/xiaohon…
13. Figma Context MCP
github.com/GLips/Figma-Conte…
14. Unity MCP
github.com/CoplayDev/unity-m…
15. Apify MCP
github.com/apify/apify-mcp-s…
Drop any of them under mcp_servers in your Hermes config and run /reload-mcp.
subhojit banerjee☦️ retweeted
Replying to @bigbasket_com
@bigbasket_com you morons have stopped allowing access to the app from outside India, how are people outside india going to order groceries for their old parents in india you fools
subhojit banerjee☦️ retweeted
Try it in your browser: udgamlabs.com/models/laghu
Run it locally (Ollama, LM Studio, llama.cpp): hf.co/udgamlabs/Laghu-GGUF
Every step, every dollar (7 min): youtu.be/NWpQv0Jn0pg
How we made it: hf.co/blog/udgamlabs/laghu
Code: github.com/udgamlabs/laghu
I train small models from scratch on small budgets and publish all of it. Follow for the next one.
subhojit banerjee☦️ retweeted
The MOST obvious (but REALLY good) STARTUP ideas right now:
1. Zestimate for company data (what your data is worth to an AI lab)
2. Google Analytics for agents (see which agents visit your site and what they want)
3. Y Combinator for 1 person companies (back solo founders running a team of agents)
4. Okta for agents (who each agent is and what it's allowed to touch)
5. Marie Kondo for your company's data (clean it up so agents can use it)
6. FICO for agents (a trust score before an agent can spend money)
7. Geico for agents (insurance for when an agent makes an expensive mistake)
8. Datadog for agents (see every step your agents took)
9. Twilio for agents (phone numbers and texting for agents)
10. Upwork for agents (hire an agent for a job instead of a freelancer)
11. ESPN for AI (every model launch covered like a big game).
12. Shutterstock for robot training videos (get paid to film yourself doing chores)
13. Credit Karma for what AI knows about you (see it, fix it, delete it)
14. Norton for AI scams (protect your parents from fake voices and emails)
15. PayPal Buyer Protection for agent purchases (fixes it when your agent buys the wrong thing)
16. Geek Squad for AI (set up personal agents for families and small businesses)
17. Dell for private AI (AI in a box for firms that won't use the cloud)
18. Carfax for used GPUs (a health report before you buy secondhand)
19. Airbnb for idle GPUs (rent out computing power you're not using)
20. Rotten Tomatoes for AI models (real reviews instead of benchmarks)
21. G2 for skills (find the best skills for your agent)
22. Dropbox for your AI memory (take what your agent knows about you to any agent)
23. Canva for physical products (design it with AI and get it made)
24. ADT for your personal agent (alerts when your agent does something it shouldn't)
25. Hertz for robots (rent one for a move, a deep clean or an event)
26. Harvard Business School for internet businesses (a real education in building online)
27. Workday for agents (onboard, manage and offboard your digital workers)
28. Getty Images for your face and voice (license your likeness to AI safely)
29. Meta Muse for older adults (and other high LTV niche)
30. Google Ads for agents (pay to be what the agent recommends)
31. Loom for agents (a walkthrough of everything your agent did)
32. Plaid for agents (approved access to banks, airlines and stores)
33. Jiffy Lube for vibe coded apps (quick fixes when your AI built app breaks)
34. Ancestry for datasets (where your training data came from)
35. The organic label for human-made content (proof a person made it)
36. Strava for builders (track what you ship every week)
37. Honey that haggles for you (your agent negotiates the price)
38. SEM Rush for personal agent platforms (like Muse etc)
39. Frequent flyer miles for agents (loyalty programs built for agents)
subhojit banerjee☦️ retweeted
50 Sites That Show What Billionaires Buy and Read for Free🤯
1.) dataroma.com → Current portfolios of famous investors
2.) whalewisdom.com → Quarterly fund trades
3.) capitoltrades.com → Stock transactions of US politicians
4.) openinsider.com → Executives' own company stocks
5.) 13f.info → Making fund filings readable
6.) sec.gov/edgar → Companies' official reports
7.) berkshirehathaway.com → Buffett's letters since 1977
8.) buffett.cnbc.com → Video archive of Buffett meetings
9.) oaktreecapital.com → Howard Marks' investor memos
10.) bridgewater.com → Research from Ray Dalio's firm
11.) principles.com → Dalio's principles for free
12.) gmo.com → Jeremy Grantham's market notes
13.) aqr.com → Quantitative fund research
14.) pages.stern.nyu.edu/~adamoda… → Damodaran's valuation data
15.) aswathdamodaran.blogspot.com → Valuation professor's blog
16.) collabfund.com/blog → Morgan Housel's money writings
17.) fs.blog → Mental models for decision-making
18.) am.jpmorgan.com → JPMorgan's free market guide
19.) ark-invest.com/research → Innovation-focused fund research
20.) a16z.com → Venture capital essays
21.) csinvesting.org → Value investing lecture notes
22.) gsb.columbia.edu/valueinvest… → Columbia's value investing archive
23.) fred.stlouisfed.org → Fed's economic data archive
24.) federalreserve.gov → Fed decisions and statements
25.) ecb.europa.eu → European Central Bank data
26.) bis.org → Reports from the central bank of central banks
27.) tradingeconomics.com → Countries' economic indicators
28.) macrotrends.net → Companies' long-term charts
29.) stockanalysis.com → Company financials for free
30.) companiesmarketcap.com → Companies' market cap rankings
31.) finviz.com → Stock screener and heat map
32.) portfoliovisualizer.com → Portfolio backtesting
33.) tradingview.com → Charts and market analysis
34.) koyfin.com → Professional market screener
35.) annualreports.com → Company activity reports archive
36.) investor.gov → SEC's investor education site
37.) bogleheads.org/wiki → Passive investing encyclopedia
38.) investopedia.com → Financial terms dictionary
39.) corporatefinanceinstitute.co… → Free finance guides
40.) cfainstitute.org → CFA research publications
41.) morningstar.com → Fund and stock evaluations
42.) simplywall.st → Companies' visual analysis
43.) valueinvestorsclub.com → Archive of selected investment theses
44.) etfdb.com → Seeing what's inside funds
45.) curvo.eu → European fund backtests
46.) ofdollarsanddata.com → Investment writings with data
47.) awealthofcommonsense.com → Market history notes
48.) theirrelevantinvestor.com → Investor psychology writings
49.) visualcapitalist.com → Explaining the economy with charts
50.) ourfiniteworld? → (see note)
The rich don't hoard information. They just don't tell you where it is.
Save it, follow @Liyana_Coder you'll need it.
Not investment advice.
@Liyana_Coder
subhojit banerjee☦️ retweeted
🚨BREAKING: The Most Comprehensive Ivermectin–Mebendazole Cancer Paper to Date Finds 37 Distinct Anticancer Mechanisms Across 25 Human Cancer Types
We synthesized DECADES of evidence from over a HUNDRED studies, revealing one of the BROADEST repertoires of anticancer mechanisms ever documented for an off-patent drug combination:
IVERMECTIN: 29 distinct anticancer mechanisms across 20 human cancer types.
MEBENDAZOLE: 26 distinct anticancer mechanisms across 17 human cancer types.
COMBINED: 37 distinct mechanism classes across 25 human cancer types in preclinical research.
Across laboratory and animal studies, ivermectin and mebendazole have demonstrated multiple overlapping and complementary anticancer mechanisms:
• Disrupt cancer-cell division
• Trigger programmed cancer-cell death
• Interfere with tumor energy production
• Inhibit tumor blood-vessel formation
• Suppress cancer-cell growth signaling
• Target cancer stem-like cells
• Interfere with metastatic spread
• Modulate antitumor immune responses
• Disrupt cancer-associated DNA repair pathways
• Counter multidrug-resistance mechanisms
• Enhance sensitivity to certain chemotherapies
• Influence pathways involved in radiation resistance
This is a REMARKABLE breadth of anticancer mechanisms for two inexpensive, off-patent drugs—spanning nearly every major hallmark of cancer biology.
The landmark 44-page scientific paper was produced by our research team at the McCullough Foundation and The Wellness Company.
Importantly, both compounds have decades of human use, established manufacturing infrastructure, well-characterized pharmacology, and low generic-drug costs.
The data can NO LONGER be ignored. The scientific case for launching large clinical trials is stronger than ever.
In the meantime, cancer patients will continue exercising their “right to try.”
subhojit banerjee☦️ retweeted
Damián tiene razón: hoy un VPS barato más una IA te reemplaza Railway. Yo tengo mis cosas self-hosted y así lo armaría, paso a paso:
1. Sacá el VPS (Hetzner, DigitalOcean o Vultr) con Ubuntu LTS y entrá con llave SSH, nunca con contraseña.
2. Antes de pedirle nada a la IA, creá un usuario sin root con sudo. Que la IA trabaje con ese usuario, no con root.
3. Primer pedido: "cerrá todo menos 22, 80 y 443, desactivá el login por contraseña e instalá fail2ban". Seguridad primero, proyecto después.
4. Segundo pedido: "levantame el proyecto en Docker Compose, con Caddy adelante para el HTTPS automático". Que todo quede en un repo: si mañana se rompe, lo volvés a levantar en minutos.
5. Backups diarios fuera del servidor. Un VPS sin backup es una apuesta, no un deploy.
6. Pedile que te explique cada comando antes de correrlo. Si no lo entendés, no lo corras.
La IA te ahorra horas. El criterio sigue siendo tuyo. 🧉
Loco, ya están grandes para seguir usando esas plataformas como Railway para deployar su web, ahora que tienen AIs que hacen absolutamente todo saquen un VPS (DigitalOcean, Hetzner, Vultr) donde quieran, le pasan la IP del VPS a su AI y le dicen configurame todo y levantame este proyecto ahi.
Listo, se ahorraron mucha plata por mes, dolores de cabeza con filtraciones y caidas.
subhojit banerjee☦️ retweeted
🚨Big Pharma is Terrified of This Man: Jim Humble Says “Chlorine Dioxide Cures 98% of All Diseases — Backed by Over 100,000 Patents!”
Did you know one simple, suppressed substance could destroy Big Pharma’s control over your health forever?
Legendary discoverer Jim Humble boldly declares:
“Chlorine Dioxide cures 98% of all diseases and is backed by more than 100K patents worldwide.”
Malaria gone in hours. Cancer under attack. Autism symptoms reversed. HIV, arthritis, diabetes, Lyme, and even “incurable” conditions — all reportedly falling to this powerful oxidizer.
They don’t want you to know about it.
But millions are waking up and taking their power back.
Have you heard of MMS Miracle Mineral Solution/ Chlorine Dioxide? Have you used it to heal illness?
subhojit banerjee☦️ retweeted
a MUST READ for founders who can build a product but can’t get anyone to buy it.
@rlakhnyuk's $25K/month playbook for making money with "boring" products in a crowded market.
P.S. both Cloudflare Tunnel and Tailscale are 100% free so you don't really have any excuses
They make money upselling to big companies and enterprise so they'll probably stay free
Maybe important to repost this now that everyone is getting hacked
If you have a VPS, I recommend you to get if off the public internet and into a Tailscale network and Cloudflare Tunnel ASAP
It doesn't matter you have SSH key login, this is about exploits and vulnerabilities AI is finding
There's 2 ways into a VPS:
- via your site/app (like the PHP exploit of this week with a file upload hack)
- via your server (like hacking into SSH or some open port)
The site/app side you put on a Cloudflare Tunnel and then with the default Free Plan it has WAF security enabled by default and they jump on new security exploits fast, it means that the only way to reach your website is via the Cloudflare Tunnel, your server does not respond to port 80 or 443 anymore
The server side you put on a Tailscale network which is like your own private network that nobody can access but your own devices (like your laptop, your phone and your VPS) kinda like a VPN
Then you switch off ANY regular access to your VPS on the firewall, in my case I use Hetzner and you create a new firewall then empty it and that blocks everything (test it)
How do Tailscale and Cloudflare Tunnel get in then? They don't. They go outbound. They make a tunnel to outside which sends and receives
Now you've reduced your attack vector to Tailscale and your Cloudflare Tunnel, much safer, not perfect btw, but nothing is
































