NOtrust

@i_am_shruti68 @sentdefender No results? Half the Iranian navy sunk. Massive portion of their missiles and launchers destroyed. The entire top and upper tier of the govt killed. Massive destruction of their capitol city and oil infrastructure.

@sentdefender bruh Iran is the chad country of this century so far (apart from killing protesters etc) they really showed you'll get bullied till you don't smack the bully in the face, usa literally spent 25$ billion dollars on this war and counting... but still no results

The U.S. Treasury Department’s Office of Foreign Assets Control has issued waivers temporarily lifting nearly all sanctions imposed on tankers originating from or carrying oil from the Islamic Republic of Iran, dubbed “Authorizing the Delivery and Sale of Crude Oil and Petroleum Products of Iranian-Origin Loaded on Vessels as of March 20, 2026” effective for vessels loaded on or before 12:01a.m. EST, March 20, 2026 through 12:01a.m. EST, April 19, 2026.

@sharbel How didn’t you realize this? It’s a win win. Security for their clients and data for their ais. everything online has always just been farming you. “How can I trick someone into exchanging their time/information/expertise for nothing in return.” A whole article??

> you open a website > it says: "click all the traffic lights" > it takes you 10 seconds > 200 million people do this every day > you weren't passing a security check > you were labeling training data for a self-driving car > google acquires recaptcha in 2009 > deploys it on every bank, store, and government site on earth > turns the entire internet into an unpaid AI annotation factory > waymo is now worth $45 billion > you didn't pass the test. you built it.

@Robkearney1981 @districtfenian My brethren! Typical Irish face!

@districtfenian It's this photo below that I colorized. If the site did any AI on it, that was unintentional on my part. I just wanted the color.

Members of the United Irish Counties Association carry their banner of resistance proclaiming "ENGLAND GET OUT OF IRELAND" for the first time at the New York City St. Patrick’s Day Parade in 1948 🇮🇪☘️🇺🇸 This banner has been carried at the parade continuously every year since then.

@DraperMP @mattyglesias If you support Palestine without supporting Irish unification you’re missing the point.

@mattyglesias That’s a perfectly acceptable answer. What did you expect him to say?

This has to be the most right-wing take on Ireland that any NYC mayor has had since the era of Bill the Butcher.

@Robkearney1981 @districtfenian My brethren! Typical Irish face!

@washghost1 You don’t need to listen to him. It’s a reality. You can’t stop the tide of time dude.

I’m from the north and I don’t even like New Yorkers moving here so I can imagine how the southerners feel

@washghost1 Notice how first it’s the immigrants “they’re” moving in to ruin the place. Now it’s just libs “they’re” moving in. Hmm.

@washghost1 How about thank you. New York. Cali. They pay all your govt budget expenses. The least you can do is accept some of them moving in to take advantage of all those hard paid taxes you guys spend on your non productive welfare recipient states. Roads sewars water. You’re 🇮🇳 without

@waefrebeorn @sudoingX Can you elaborate. Say you have 128 gb ddr4 and 3070ti. Is there a way to layer manage that to run a decent model?

@sudoingX none of you understand layer management and running the higher layers in vram and lower layers in sysram a single 5050 with 8gb vram can load with a 96gb sysram and token max via layer management reaching similar speeds ONLY ON MoE MODELS WITH 8b

drop your GPU below. i'll tell you exactly what model and config to run on it. here's what i've tested and verified on real hardware: RTX 3060 12GB - Qwen 3.5 9B Q4 - 50 tok/s - 128K context RTX 3090 24GB - Qwen 3.5 27B Q4 - 35 tok/s - 300K context RTX 3090 24GB - Qwen 3.5 35B MoE Q4 - 112 tok/s - 262K context 2x RTX 3090 - Qwen3-Coder 80B Q4 - 46 tok/s - full VRAM all running llama.cpp with flash attention. every number is real. every config is tested. if your card isn't on this list drop it below and i'll tell you what fits.

@cjzafir What a dumbass list

Some secrets: 1. Opus 4.5 > Opus 4.6 2. Codex 5.4 1M context is actually 512k 3. Qwen 3.5 is best OS model series (RN) 4. MCPs are dead 5. Openclaw was a prototype 6. Apple is the new NVIDIA 7. Perplexity is biggest grifter 8. Paper. design > figma 9. Ollama is just a crappy wrapper 10. Open source AI is ahead already.

@cjzafir @javbuitrago Explain why any of those except Mac studios are at all good for ai focused consumers??? They have no vram and can barely run top os models. Not to mention os models are light years behind frontier models.

@javbuitrago Apple chipset for consumer hardware is a big opportunity for them. (macbooks, mac minis, mac studios and they will definitely make AI focused hardware too)

@CaptainCons They’re birds.

The best players in the world are playing gorgeous TPC Sawgrass for the esteemed Players Championship Bryson is battling bugs the size of your fist and falling into bunkers in front of zero crowd

@25THPRMR @sharbel You absolutely do not need a Mac mini. Enormous fomo trend that adds no value. No serious models can run locally on a Mac mini. You should start with Claude code or antigravity which has a free tier.

@sharbel fair enough

the fastest growing GitHub projects this month: 1. openclaw/openclaw (122K stars) your own personal AI assistant, runs 24/7 on any OS (what I use to run all my agents) 2. obra/superpowers (30.7K stars) agentic skills framework. plug-and-play tools for AI agents 3. ruvnet/RuView (30.4K stars) turns regular WiFi signals into real-time human pose detection no cameras. no sensors. just WiFi. 4. 666ghj/MiroFish (17K stars) swarm intelligence engine that predicts anything 5. moeru-ai/airi (16K stars) self-hosted AI companion with real-time voice chat runs on your own machine. you own it. 6. shanraisshan/claude-code-best-practice (11.8K stars) the best practices repo for building with Claude Code 7. badlogic/pi-mono (11.8K stars) full AI agent toolkit: CLI, unified LLM API, web UI, Slack bot 8. bytedance/deer-flow (10.4K stars) ByteDance's open-source SuperAgent. researches, codes, creates on its own 9. shareAI-lab/learn-claude-code (9K stars) build a Claude Code clone from scratch. bash is all you need. 10. p-e-w/heretic (7.6K stars) removes guardrails from any language model automatically the pace of AI right now is insane. bookmark this. next month's list will look completely different.

@jluan They’re not wrong lol what’s weird about this?

apparently this is what the chinese AI ecosystem thinks our grand american AI master plan is! (this is not a joke! forwarded to me by an attendee at a real chinese ai conference!)

@713Capital @Zeneca You don’t think openclaw was vibe coded? Also they’re all just python networking. Open claw is just a prebuilt one. Nothing so special about open claw. You can easily vibe code your own python program to route the agents.

@Zeneca One is a vibe coded clone of the other. That’s all I’ll leave you with. Tread carefully

thoughts on openclaw vs hermes? has anyone switched from one to the other, successfully bringing their memory systems and other things?

@SystemDaddyAi @RoundtableSpace No way this works as well as opus sorry.

@RoundtableSpace autonomous agent in 5MB of RAM built by Harvard and MIT students. meanwhile enterprise AI teams spending $50K/month on cloud compute to do the same thing. complexity was never the requirement. it was the business model. 💀

YOU CAN NOW RUN A COMPLETELY AUTONOMOUS AGENT WITH LESS THAN 5MB OF RAM - Fully modular: models, tools, memory, and channels can all be swapped out easily - Connects directly to 15+ platforms — no cloud server rental required - All your data stays 100% private and never leaves your device - Completely open source — 100% transparent and community-driven

@libriscent You can’t research for weeks and buy impulsively that’s inherently contradictory. Even if you do a ton of research and the. Choose something randomly that’s not on impulse you e already done the work that prevents it from being acting on impulse.

Autism: Researches every purchase for weeks. ADHD: Impulsive shopper. AuDHD: Researches for weeks, buys it impulsively anyway.

@klouttf @Krysta92kat_ @SomaKazima2 Ty for the correction/specifics 🙏

@proofNOTtrust @Krysta92kat_ @SomaKazima2 these are not box braids their fulani braids box braids do not braid against the scalp these are box braids

Since y'all don't like to listen to us, listen to one of your own 😭😭😭

@okaythenfuture @Mil2Roro Absolutely insane take. We have more missiles and way more ships. And our radars and tech actually work and aren’t temu crap.

@Mil2Roro You aren't the industrial superpower this time little bud, this time its China that will have the Mala naval ships, no ice cream for American sailors this go around.

Americans don't really get that its all over. The American military mystique is effectively DEAD. Just like Russia's died in 2022 in Ukraine, America's has died in 2026 on its 250th anniversary in Iran. The first and second island Chain are FINISHED. If Iran can do this much damage to Gulf + Middle Eastern sites, everywhere from Guam to Okinawa is finished. They stand no chance against Chinese Ballistic missiles and drones. Its over, the empire is gone and it died this week. You can make your peace with that, or cope.