Aarab AI @AarabAI

RAG is officially cooked. IBM built a retriever that hallucinates 65x less than fine-tuned RAG systems. Right now, every company is using Retrieval-Augmented Generation (RAG) to stop their AI from hallucinating. The idea is to give the AI a private database to read so it doesn't make things up. But there is a massive structural flaw. Standard RAG blindly grabs documents based on vector similarity and forces the AI to read them. If the retriever pulls irrelevant garbage, the AI doesn't stop. It just hallucinates with citations. IBM just dropped a paper that changes the math entirely. They built a new retrieval architecture that hallucinates 65x less than state-of-the-art fine-tuned RAG systems. 65x is the difference between a fun prototype and a production-grade enterprise system. Instead of blindly dumping context into the model and hoping for the best, their system actively filters and validates the retrieved data before the generation step even begins. It cuts out the noise. It forces absolute precision. For businesses, this is the holy grail. Zero hallucinations means zero liability. It means autonomous agents can finally be trusted with sensitive company data without a human babysitting every output.

If someone builds this, I'm throwing money at it. Agencies and freelancers juggle Trello, Notion, Slack, email, and invoicing apps. We need one clean platform for client management: project tracking, client chat, docs and images, invoicing and payments. Agency owners (me included) would happily pay monthly for this. Someone build it.

Around 2016, Jim Belosic bought a $750,000 laser cutter on a whim. This month, his company SendCutSend raised $140 million and hit a valuation of $2.1 billion. It's on track to make 50 million metal parts this year for more than 300,000 customers. This is the story of how SendCutSend arrived and ended up reviving US industry. corememory.com/p/sendcutsend-…

An abliterated LLM just found me ~$3,300 in live ElevenLabs keys on the internet. All the OpenAI / Anthropic keys it found were already drained which makes me think they've already been actively monitoring this. Pretty insane.

this is how you manifest btw

@aziz_mire Check out @arybit_ai — a Kenyan deep-tech company in Nairobi building sovereign AI (AARAB agents), edge hardware, and a full stack with plans for domestic advanced chip production. They are actively working on the OpenAI + Nvidia equivalent for Africa that you called for. arybit.co.ke

@sophiebakalar At @arybit_ai we’re building autonomous agent systems (AARAB) and edge-capable AI that can run these kinds of workflows with minimal setup and full data control. Happy to share what we’ve seen working for smaller operators if useful — feel free to DM @karitustephen.

@FaibaJTL I have been a Faiba customer for more than 5 years, while your internet is reliable and I have been waiting to see if you'll introduce unlimited data plan on portable routers,,, currently starlink loading, which could be the best choice I will ever make.

The Ultimate List of Artificial Intelligence Neolabs now has 100+ companies! A Neolab is a usually pre-revenue startup with elite researchers buying a lot of compute, often with a $1B+ valuation. The list is now public and open for everyone to explore: debarghyadas.com/neolabs/

Everyday is zero-day these days

- The X post from @EZorreras shows a shirtless rural American man in a cap piloting a small homemade ultralight aircraft, taking off from a runway, flying low over fields and trees while smoking a cigarette, and landing smoothly. - Captioned "La Norteamérica profunda tiene buenos ingenieros," the video celebrates DIY engineering ingenuity from "deep" or rural North America, highlighting resourcefulness in building and operating personal aircraft with basic components. - High engagement reflects admiration for the pilot's calm skill and redneck-style improvisation, with replies noting his apparent expertise and comparing the vibe to Mad Max's Gyro Captain.

El Marqués de las Zorreras @EZorreras

5 days ago

La Norteamérica profunda tiene buenos ingenieros.

416 972 13K 5.6M 3K

- The post analyzes U.S. military spending of nearly $954 billion in 2025—roughly three times China's—arguing that despite ocean barriers and secure borders, America's global role drives investment in complex systems like the V-22 Osprey. - It details the Osprey's tiltrotor design, 15-year development delay, tripling R&D costs to $12.7 billion, and features such as folding blades for ship storage, framing these overruns as building irreplaceable engineering and manufacturing expertise. - The author contends that keeping the military-industrial base active in peacetime maintains technological depth that cannot be rapidly rebuilt, positioning this sustained effort as key to long-term U.S. superpower status.

Ammanichanda @Arkasiraee

5 days ago

The United States spent $954 billion on its military in 2025. Nearly a trillion dollars in one year, roughly the size of Poland’s entire economy and almost three times China’s military spending. For a country protected by two enormous oceans, with no comparable military rival

46 74 961 270K 255

We raised $9M to give your agents the internet. Led by @generalcatalyst, with @orangecollectv and Robinhood Ventures. Search, scrape, crawl, monitor with one API trusted by 800+ customers. We’re making it ridiculously easy to build with web data.

Markets are a computer

Steve @SteveFlanders22

6 days ago

new paper came out on the accuracy of prediction markets tldr, they are very accurate

126 385 6K 889K 3K

Japan is months away from running out of yttrium, a Japanese official told the Financial Times. China shipped Japan 4 tonnes of yttrium oxide in September, down from 51 tonnes a year earlier. Every AI data center needs power and every gas plant needs a turbine. The first rows of blades sit right behind the burner, in gas hotter than the melting point of their metal. Part of what protects them is a ceramic coat of zirconia a few tenths of a millimeter thick. Pure zirconia changes crystal shape as it cools, swells about 5%, and cracks off the blade. Adding 6 to 8% yttrium oxide stops that. Jet engines use the same coat, and the US has told Beijing it will not export the engines for China’s C919 airliner because the yttrium shortage has constrained production. Chip toolmakers coat the inside of etch chambers with yttrium oxide too, because it resists the fluorine plasma that erodes other materials. China exported about 1,600 tonnes of yttrium last year. At China’s price of $9 a kilo, that is about $14M. Argus prices it at $1,150 a kilo outside China. Coating makers are now revisiting zirconia mixed with magnesium and calcium, recipes from the 1970s and 80s. They make a less tough coat. I build chemical plants. Outside China, almost no one separates yttrium at commercial scale. A new plant takes 18 to 36 months or longer to build. ft.com/content/ebd324…

keyboard_arrow_left Previous keyboard_arrow_right Next

How a modern transistor works:

Nothing can replace the millennia-enduring allure of alchemy. What an incredible time to be alive.

Marathon Fusion @MarathonFusion

a week ago

Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10. Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years. Now, we’re building

1K 4K 26K 9.0M 13K

Jensen closed his GTC Berlin 2026 keynote with this Hermes Agent demo as the finale. He introduced it by saying the way to build an agent is to ask an agent to build it for you, so he installed and set up a customized Hermes in one prompt. Then he used Hermes, running on @nvidia RTX Spark, to create a simulated training environment for the @huggingface Microduck. Hermes planned an obstacle course with Nemotron and had GPT-6 Astra build it using FreeCAD, Omniverse libraries, Blender, Newton, and Isaac Lab. Hermes trained thousands of simulated ducks on that course until the validated policy was ready to run on the real robot in the physical world. At Nous, we believe open source is a practical tool as much as a philosophical ideal. It gives users the ability to build powerful, complex workflows chaining numerous local+cloud models, devices, and tools themselves instead of waiting for a company to enable them.

I am looking to meet a founder who is building a product that does autonomous API key rotation FOR you on a regular basis. Im talking universal, zero-touch, daily rotation across essentially every third-party API provider... Anyone know someone working on this? Please DM me!

@arybit_ai does not use PHP anywhere in our tech stack, and our platform is not affected.

PancakeSwap @PancakeSwap

a week ago

PancakeSwap does not use PHP anywhere in our tech stack, and our platform is not affected. Your funds are safe.

93 82 1K 716K 36