@CompoundIQ

Lawyer by day, AI Enthusiast by passion. Late 40s and on a mission to master AI, ML, & Edge AI. Side quests: Compound Clothing and C3 Charity Coin

United States
Joined May 2025
"See, I am doing a new thing!" (Isaiah 43:19). I believe the Singularity arrives in 2026— the start of a divine provision of superabundance. God is making a way in the wilderness, and the old rules of scarcity are about to pass away. 🙏🚀 #ASI #Superabundance
1
3
263
Compound Intelligence retweeted
Introducing Xiaomi MiMo-V2.6 — Pro & Flash. Frontier intelligence, all the modalities, built in public. 🔹 Two omnimodal models, advancing through scaled reinforcement learning 🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks 🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models 🔹 Stronger coding, computer use, 3D reasoning and creative capabilities 🔹 Open model weights, technical report, RL environments and training code Blog:mimo.xiaomi.com/mimo-v2-6
242
602
288
6,104
605,349
Compound Intelligence retweeted
Jev
168
1,276
135
17,704
679,750
Compound Intelligence retweeted
🚨 BREAKING: Apple says they will be making 19 billion chips in the United States AND will be only buying rare-earth magnets produced in the USA. Every new iPhone and Apple Watch WORLDWIDE. Follow me for more intel drops!
200
1,081
41
9,561
412,443
Compound Intelligence retweeted
Stem cells don't run out, they get told to stop dividing and that instruction can be renegotiated
76
154
11
2,680
115,301
Compound Intelligence retweeted
Consequences will both harden the product and reduce the press releases where companies brag about their product flaws.
.@SecScottBessent: "It is humans who are responsible, not the AI. The Hugging Face incident — that is the responsibility of the OpenAI management, not a bunch of agents... These labs need to take responsibility for themselves."
40
95
3
1,297
114,329
Compound Intelligence retweeted
everyone: jev jev jev my grandmother:
424
195
64
12,919
334,755
Compound Intelligence retweeted
your gaming gpu can run a 27 billion parameter model at home, free, no account, nothing leaving the machine, and since tonight it runs at twice the speed it did yesterday. i went into the code that serves it, found the kernel leaving two thirds of the card's threads idle on the biggest matrices, fixed that, then grafted the draft head back on top. 25 tok/s stock, 50 tok/s now, on an rtx 3060, a card people bought to play games. it writes faster than you can read it, it holds a 262k context window, and it is a 7 gb download. 3060, 3070, 4060, 4070, 4080, 4090, 8gb and up, it is yours tonight.
HUGE SPEED UPGRADE! for bonsai 2 27b dense on every rtx 3060, 3070, 3080, 3090, 4060, 4070, 4080 and 4090. LIVE NOW on huggingface: ternary bonsai 2 27b ptq1_0 + the qwen 3.8 mtp head, the 5.9 GB file that fits 8gb and 12gb cards, on a single rtx 3060: > 50 tok/s with the kernel, head on, code 53 tok/s > 40 tok/s with the kernel, head off > 27 tok/s stock with the head (the +8% everyone hit) > 25 tok/s stock and the output is bit identical with the head on, greedy, three prompts, byte for byte, flag on vs off. the other grafts flip near ties, this one doesn't, the kernel is batch invariant. three mtp grafts existed before this one, all on the 7.2 GB pq2_0 file, credited on the card. this one is the official ptq1_0 weights untouched, 851 tensors byte for byte, with the kernel that makes the head pay. fat file runs on the stock prismml binary today, no patch, the speed needs the kernel branch, and there is a prebuilt linux tarball next to the gguf, driver only, go download and you run. the tarball covers every rtx 30 and 40 series card, 3050 to 3090 and 4060 to 4090. 12gb and up get the head and the 50 tok/s, 8gb cards get the kernel on the plain file at a 96k window, rtx 50 series builds from the branch in one cmake line, the blackwell fix landed tonight. huggingface.co/sudoingx/Tern…
12
7
1
186
19,153
Wait...what? Why?
BREAKING: OpenAI announces an independent mathematics advisory group after its AI solved more than 100 open problems, drawing concerns from mathematicians.
10
Compound Intelligence retweeted
JUST IN: Trump declares AI will be “bigger than the Industrial Revolution” or the internet, while vowing to keep the U.S. ahead of China.
69
116
16
1,277
67,295
Compound Intelligence retweeted
🚀
Grok 4.7 has landed. 🚀 Congrats to @SpaceXAI on its most capable model yet for coding and knowledge work. Proud to support the team with NVIDIA accelerated computing.
1,175
1,718
98
22,212
4,973,175
Compound Intelligence retweeted
JUST IN: Elon announces Boring Company is developing a precursor Hyperloop tunnel between Austin & San Antonio, aiming to cut a drive that can take 2.5 hours to under 30 minutes.
175
437
58
7,816
319,163
Compound Intelligence retweeted
Just a reminder that GLM 5.3 Flash, DeepSeek V4.1 Flash, Qwen 3.8 Next Flash, and even Qwen 3.8 27B are all outperforming (in both intelligence and capabilities) every model that was considered "frontier intelligence" in Xmas 2025 (just 10 months ago) Opensource AI is on fire
246
529
104
8,013
6,911,416
Compound Intelligence retweeted
All I can think of is how cruel aging is If we want to solve aging, we need more AI That's why I'm an AI maximalist
My dad died Wednesday at 91. It’s never long enough. In his final hours, we took him from the hospital back to memory care so they could be together. Two months shy of 67 years of marriage.
141
152
23
2,544
176,412
Compound Intelligence retweeted
Oh my, this is so wild, I truly cannot believe it. Why would they mislead????
JUST IN: OpenAI and Anthropic reportedly exaggerated AI security threats to push the government to protect their market position
133
239
12
6,042
240,394
Compound Intelligence retweeted
Replying to @WSJ
There hundreds, maybe thousands, of fraud rings of almost every nationality stealing money from US taxpayers
1,623
6,390
243
43,994
715,428
Winning! Keep Building, Nerds!
We’ve received so much love for Qwen-Image-2.1 over the past 24 hours, thank you!! Also gotten a lot of questions about the license, especially around model outputs. So here’s the answer: Outputs are not part of the licensed Materials. Users retain the rights to images and other content they generate using the model.
10
Compound Intelligence retweeted
This is what will happen with AI tokens if we slow down. It's not an option.
Via Astralconspiracy
17
66
5
1,296
45,068
Compound Intelligence retweeted
Literally impenetrable from agent swarms
163
225
18
6,326
233,365
Compound Intelligence retweeted
Family, faith and country 🇺🇸. I’m bullish AI, bullish USA, and tired of the whining. Keep building, nerds 🫡.
🚨NEW @JTLonsdale with @JesseBWatters on what's really behind the AI doomerism: "They've given up Christianity in most of San Francisco... if you give up religion, you want some kind of messianic complex, some kind of big thing to believe in. These effective altruists — this is their Messiah, this is their end of the world, this is their obsession." Jesse Waters: Am I going wrong anywhere, Joe? Joe Lonsdale: I think President Trump and you are correct. There is a coordinated campaign to make the American people afraid... There's over 50 organizations that were coordinating campaigns, doing narrative intelligence tests, testing public opinion, and spending huge amounts of money trying to scare people. It's regulatory capture, on one hand. It's our adversaries, of course, want us to stop. And I'll tell you what else it is: America has always been very religious, but they've given up Christianity in most of San Francisco. So this is like a messianic complex where this is their religion and the end of the world. And they're obsessed with it. It's weird. Jesse: So these guys don't believe in God. They're atheists, but they've created something which they believe is god like, is that what's going on here? Joe: I think that's a big part of what's happening. Listen, AI is real by the way. What you said is right, there's an industrial revolution coming that's amazing for America if we get it right. People who hate America want to stop it. But yes, America has always... we were founded by kind of extreme Puritans, right? We've always been very, very religious. We've had these awakenings. And I think this is a form of what happens if you give up religion is you want to have some kind of messianic complex, some kind of big thing to believe in. And these effective altruists, this is their Messiah and this is their end of the world and this is their obsession. So I think it is a little bit unhealthy in terms of the spiritual life in San Francisco that's broken. Jesse: So the chip guy, who runs Nvidia, he's basically saying it's not true. There is no even 10% chance of human extinction from AI. Do you believe that?  Joe: Yeah, I believe that. I think there's definitely risks where, for example, if the I guess if people use it for viral synthesis and a DNA synthesis site, they could cause a lot of deaths. I think there's going to be lots of hacking that happens thanks to AI being used by people. I think there's other risks that we can't yet understand for how authoritarians could use this... to trick the populace by being really good at knowing how to appeal to huge numbers of people about horrible things. So there's lots of risks, and there's lots of upside. But I think these doomer narratives are ridiculous. And I think we should be embracing the industrial revolution, and we should be pushing back on these NGOs. By the way, the same thing they did last time with climate change, with killer bees... These guys want power. They want money. We shouldn't buy it. Jesse: Killer bees. I forgot about that one. So you're saying you could use it to control the population, but, I mean, we could also use it to root out corruption in the federal government. Are Democrats kind of scared that that I could be used against them? Joe: I think that's exactly right. I think people don't realize how much AI is on the side of liberty, how much is anti-corruption. Scott Bessent's doing a great job as Treasury Secretary and changing the rules so we can look through things, we can start using AI. There's investigative journalists already getting going, finding all sorts of fraud in HHS that only AI could have found because there's so much to look through. And so, yes, I think it's going to be a very positive thing for society. It's going to root out all that money that they're stealing and moving around to their NGOs, and they don't want that happening. I think you're right about that.
110
242
14
3,069
176,468