Thanks to @attilavaszka at Quarter Digital for the time and design help on ArcLoom’s homepage. Come see the refreshed site arcloominfra.com/ and his work at quarter.digital/

Gates calls AI "the most dangerous thing that humans have ever gone near" (he left Microsoft's board; Microsoft didn't leave OpenAI). How can you both throw billions of dollars behind an initiative and then go online and decry it like an old man crying, "Stay off my lawn"? youtube.com/shorts/ubtWvzfZu…

Im confused. Did something happen with human genetics that we missed? Or maybe its just not cool anymore? Someone please explain....

AGENT WARS We asked 5 top AI models to design, engineer and 3D print a bridge to with the goal of holding the most possible weight. 🇺🇸 Anthropic: Claude Opus 5.5 (High) 🇨🇳 Kimi: Kimi K3 (High) 🇺🇸 Meta: Muse Spark 1.3 (High) 🇺🇸 OpenAI: GPT 6 Astra (High) 🇺🇸 SpaceXAI: Grok 4.7 (High) Each bridge had specific parameters. It had to span 2 feet, use a maximum 500 grams of filament, and it had to be under 18 hours of printing time. We specified what type of weight and where we would place it on the bridge. The Score: • Claude Opus 5.5 held an estimated 130 lbs! It took 9 hours and 11 minutes to print 17 parts and used 441 grams of filament. • Meta Muse Spark 1.3 held 26.5 lbs. It took 13 hours and 12 minutes to print 49 parts and used 478 grams of filament. • OpenAI GPT-6 Astra held an estimated 17.5 lbs. It took 15 hours and 44 minutes to print 29 parts and used 442 grams of filament. • SpaceXAI Grok 4.7 did not finish. On assembly it could not stand up by itself. It took 12 hours and 22 minutes to print 29 parts and used 460 grams of filament. • Kimi K3 did not finish. It wasn’t engineered correctly so the bridge could not even be assembled. It took 11 hours and 19 minutes to print 35 parts and used 446 grams of filament. That means our champion for this episode is Claude Opus 5.5. Not only did it hold the most weight (by FAR) it also took the least time to print, used the least amount of parts and used the least amount of filament. This was episode 02 of Agent Wars, stay tuned for episode 03.

And somehow this makes more sense than reality....

Two things that are crazy to see. 1. No fable usage today, didn't miss it. 2. Solid 8-9 hours of work in 4 or 5 threads, and 8% usage. I'm so looking forward to the rest of the 5.5 family. The recipe is in there somewhere @AnthropicAI, keep cooking!!

You know what I didn't do today? I didn't micromanage AI(SI) platforms, frantically bouncing back and form between Codex and Claude. I just worked for a solid 9 hours in Claude desktop, just like the good ol' days. And used..... 6% of my weekly usage.

We'll see how things look when the 24hr new model grace period wears off. I'm hopeful, Anthropic was like my AI first love, and I have struggled with the relationship the last few months. Opus 5 and I fought, like a lot. Fable was there to pick up the slack, but the plan limits left tokens in the tank. Then lately the politics of Anthropic churn my stomach and I find myself in a quandary of where I spend my dollars. But where you going to go? Benchmarks are one thing, if they can be believed. But more and more the personality (I can't think of a better way to say it and I loathe anthropomorphizing weights and math) matter more than the raw intelligence. Hoping for a solid Opus! MOGA (Make Opus Great Again)

Wait what. This is AI right? Right?....

I used AI to explain the AI pacing drama, with fruit.

Besides, it'll take way more compute.

Oh Astra...

Thats what I'm saying... These SF delulus are begging to be punished. @USDOJ start fining some corps for hacking already. Failure to do so creates the impression of lawlessness, and AI immunity.

I can’t believe the crossovers happening. But more of this, please. Stop mystifying the technology. Stop thinking through sci fi tropes and be particularly skeptical about anthropomorphizations like “agents”’and “swarms.” The technical reality is far more mundane. Try to remember these are companies developing products and services. We have laws and institutions to restrain or punish them for fraud, property damage, deceptions, etc.

There is deep wisdom in this. Foundational and core level wisdom. Whatever your belief system, gratitude to something or someone has to be part of it.

Agreed.

Wait. What.... And then you just go in to work because like, Wednesday?

Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.

OpenAI put out how its own researchers use agents, and I was a little shocked. Median researcher spend went from about zero (what?) in February to over $600 a day by August. The biggest ramp came when their new big boy model hit the scene. If anybody should already be burning tokens like this, it is the labs, and even they only just hit the gas. Three things that stuck: 1. Each researcher is running four jobs (three of them never take lunch or a smoke break). 2. Agents alone draw about two homes of electricity around the clock. 3. Everything built so far barely covers a handful of companies' token usage. How much more is going to be required before the normies start burning tokens like the labs? I think we're still so early....

Foreign dollars, focused on your feed and in your town! AI is the product and the data center is the factory so if you can't picket software, picket the factory. OpenAI recently banned two Chinese operations, one posting that data centers raise your power bill, the other attacking American AI itself. A Russian op forged fake TechCrunch articles against one. By one count, five donors in Switzerland, the UK and Denmark (the Swiss, famously neutral) put $39M into a dozen national groups running the fights in 12 states. X flagged about 200,000 suspect Chinese accounts in August, a slice of them on the data center story. State media in China, Russia and Iran all ran it, China Daily's headline was "AI boom sends electricity bills in US skyrocketing." Outside money is going into anti-AI and anti-data-center campaigns to sow discontent and fan fear. That is the tricky side of the First Amendment we all hold dear. Be mindful of who's holding the microphone, or keyboard, that is the citizen's job, and nothing gets taken at face value. Fortunately, it's easier than ever to do some discovery.