Open ai is making the play to be the fast food of the SI industry. It is what it is.
ChatGPT has brought in Messi as an ambassador.
Its official OpenAI is giving up on power users and pivoting to the novice user.
I don’t see Anthropic paying professional athletes to shill their products but rather they let their models do the talking.
OpenAI has Messi, Claude has Opus 5.5
Do you prefer Messi or Opus 5.5?
Doing great things for local models.
AgentionAI models have passed 1.1M downloads on Hugging Face. 🎉
Qwen3.8-27B AP quants at 650k+, and Gyro, my rotor quant for running Qwen3.8 Flash Next (126B) on a 32 GB card, is past 170k in under two weeks, with Strata support live.
Thank you to everyone who tested, reported bugs, shared results and supported me. Special thanks to @coldniko, who gave the first donation to the GPU fund and added Strata support.
Link in bio if you’d like to help with the next big one.
We need to get back to being leaders in production, and real prosperity.
Replying to @NathanJBarnes
America built 10,000 miles of interstate between 1956 and 1960. France built 42 nuclear reactors in the 1980s. The West used to be able to do this. And we could do it again.
HEY @X, help me find my people.
I'm building with AI every day: Claude orchestrating, GPT reviewing every change, local models on a 118GB AMD lab, Omarchy, Ubuntu, and Windows.
Into agents, local models or Linux? Say hi. I follow back real builders.
Buckle up, bumpy ride guaranteed.
wow this is bad!
ENTIRETY of PingCap database used by linkedin, flipkart, airbnb etc. has been copied and stolen due to customer credentials being stored directly in support tickets (LOL)
Attackers ran exfiltration directly on their cloud (LOL) pipeline ran for 22 days, exfiltrating data every 6 hours and NO ONE noticed. Worse, PingCap denies all of it.
Average SRE salary is $300,000 at PingCap btw
If you ever wanted to build some electronics, the doors have opened.
Replying to @Compilethings
I sort of went one better with the wrapper github.com/DingoOz/WholeBoar… It will do all the steps
My first PCB is designed and ready to be made. It’s a carrier board for the Latte Panda MU Ultra. Working on the keyboard, I/O board and antenna board. When those are finished we order and see what happens. Thank you @VectorCrossProd for TraceMaker!
@LattePandaCN @dfrobotcn @PCBWayOfficial
working on a project, going to build it live on X, If you guys would like to help out, I would be very grateful.
This is the first of many. It will be open source.
This guy knows what's up. Every home should have as many on file that can run on their hardware. Prepare for defense. We are moving at warp speed.
I am not at this level, yet. I run a 104 GB Qwen3.8 Flash-Next (4-bit) on one home server with 3 Radeon AI PRO R9700 cards, served only on my home network. It does 74 tokens/sec warm with a 131K context. I reach it through Hermes from my laptop for family documents that must never leave the house.
A 104 GB AI model running at home. Qwen3.8 Flash-Next on 3 AMD R9700 cards in one home server: 74 tokens/sec, 131K context, reachable only on the home network. "My family's documents never leave the house." How do you deal with privacy?
Opus 5.5 is still one shotting my work 86-93% and usage is still great. I don't want to your Nerf stories.
Lets GO!
New in WholeBoard: a built-in tour. 23 steps from describing your board in plain words, through schematic + ERC, ngspice sim, TraceMaker routing, DRC, 3D renders and fab files, to a JLCPCB quote and an enclosure. Which step should I show in depth?
github.com/DingoOz/WholeBoar…
Well there you go! They did the thing!
I told you bio tech is where these companies will hit the big money.
GPT-Rosalind surfaced a new lead for drug discovery: could restoring cholesterol transport protect neurons from degeneration?
DCTN1 is a gene linked to Perry syndrome, a rare neurodegenerative disorder. In mouse-brain data, disrupting it increased expression of genes involved in making cholesterol. GPT-Rosalind suggested a possible explanation: cholesterol may not be reaching where neurons need it.
That gives scientists a specific mechanism to test and, if confirmed, a direction for developing new treatments for neurodegenerative disorders.
With @perturbai_tx and @scrippsresearch, a team of thousands of GPT-Rosalind agents worked in parallel, connecting gene-expression results from millions of mouse brain cells with scientific literature in just 32 minutes. We’re now expanding this collaboration to identify new drug targets.
openai.com/index/scaling-ai-…
Local is getting very interesting these days, I need to get this one downloaded.
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from scratch in under 4 months
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
Converted my wife to Omarchy today. She was complaining about her old widows laptop and how slow it was, feels like a new machine now.












