Pinned Tweet
Maybe this is the real answer to the Fermi Paradox: every civilization eventually votes their way into total socialism, and dies out.
I wonder how long until a completely AI-made film gets widely played in major theaters. That is, not a one-off, or a dozen select theaters, but theaters nationwide / worldwide.
On one hand, I don't think we're all that close. On the other hand, I can see it happening, and "not all that close" at the rate AI is advancing might be 10 years, not 50. And more reasonably, good coherent long-cut movies will start happening within 5 years, and it'll take 5 more years after that to convince theaters that they can make money selling tickets to people to watch them on a big screen.
29
Watching tons of Chinese language posters manifest to proclaim $nvda doom just makes it even more obvious that this was all a CCP plot. Probably another leg of their campaign against Taiwan (and western AI). Xi Jinping must be desperate.
绝大多数人对黄仁勋在英伟达作用的认识还是太幼稚了。
美国那套空降兵制度(职业经理人制度),
根本不可能玩转这种高科技企业。
加拿大北电危机了,叫来一个经理人,结果北电不存在了。
黑莓危机了,叫来一个经理人,谁还知道黑莓啊。
诺基亚危机了,叫来一个经理人,诺基亚手机很快就消失了。
这种例子数不胜数。
苏妈管不了英伟达,库克管不了英伟达。
英伟达42000人,单人利润286万美元,是苹果的四倍多。
这些都是老黄的高效管理模式导致的。
2
207
#coding #skills #agents:
One thing you should be doing with your agents, where applicable: turning skills into deterministic features.
I spent a good while developing a skill that helps debug a particular setup. It knows about the code, the build, how to access the server, what processes are running, how to look at the logs, etc. There is a ton of "tribal knowledge" in there, such that even rank and file QA testers are using it now to debug issues.
But it was just a skill. The entire thing was agentic.
So, I had the agent spend time turning its own skill into a deterministic program that can do quite a lot of the work up front, with no agent. Basically just a python program that shells in, grabs screens, copies over logs, and parses the logs for a variety of known issues. Like THIS log needs to say THIS if it's healthy but if it says THIS, it means this other thing is offline. If that other thing is offline then look at THIS and see if you see any of these symptoms....
In the end, it created a deterministic program that can find all kinds of problems with no agent at all.
But I did keep the skill, and the agent. That deterministic program is just one of its tools now. It runs that first, and from that, it can dig deeper if it needs to, or just tell you what's wrong. It greatly reduced the cost of running the skill, plus it's way faster for most things. (And if we find something the deterministic skill didn't cover, we can usually add it as another thing to check for.)
1
127
Conan the Barbarian retweeted
Replying to @reddit_lies
If Reddit, Facebook, Instagram or any of the others ever revealed VPN or appstore location like X does, I think a lot of people would be in for a hell of a shock.
Starting with their advertisers. Which is why they won't do it.
Replying to @RandomPerson242 @anneapplebaum
Best estimate from OpenSecrets: Democrats accounted for about $5.52 billion of the record $11.9 billion spent in the 2026 midterms (projected total, including candidates, parties, and outside groups). Democratic candidates alone outspent Republicans roughly $2.2B to $1.4B.
I think we're in a real race: AI and tech is racing to bring about the post-scarcity economic dream, but they have to get there before the radical left destroys them, and us, with a cultural disaster that never had any chance of working out well.
2
130
Things I hate: there is a 'bloop' noise. I have no idea which app sent it. I cycle through them all and nothing appears to be new. Later, there will be another 'bloop' noise.
86
Kinda thinking today how AI makes me feel like a kid again, in the best of ways.
When I was a wee lad, I loved programming and always had disks full of half-baked projects. Another day, another idea. Driving a project to completion was never the point. I worked on it until some other idea sounded like more fun.
Then I made programming my career, spent all day Working For The Man, and this sucked the life out of my desire to go home and program for fun.
But with AI doing the actual drudgery (including, to a large degree, playtesting), the old fun is back. "Programming" was never "the old fun", to be clear. The old fun was having an idea and seeing it come to life. I again have multiple projects going on at the same time -- a somewhat dizzying 'parallel processing' mental mode that I used to do more often. Instead of making me feel mentally lazy, AI is making me feel younger and mentally active again.
As a bonus, I can now also actually ship things because that last 10% that was the least fun of all -- actually getting something finished -- can be largely done for me.
Took me like 4 tries to get through this message because I am bouncing between: the legal ramifications of my camera app; the Roblox game I am making and having the AI playtest in the background; a discussion with some old gaming buddies about whether we should turn WOW or Empyrion into a Bannerlord type game.
I also need to remember to find a time to reboot because the "PC Agent" has informed me that the fan noise I keep hearing is some kind of conflict between the BIOS, Asus, and a third program all messing with temperature settings and one of them is causing a problem.
1
97
I love how Euros think it's a sick burn to say that they are finally going to get off their asses and quit being so dependent on American power.
Like telling your wife that, you know what, I'm going to make my own sandwich!
The clock is ticking, America.
Your tech isn't irreplaceable: one by one, Europeans are building their own alternatives.
🇺🇦 Ukraine and 🇩🇰 Denmark just put the first test satellite of a sovereign Ukrainian comms network into orbit.
Less Starlink, more European autonomy. 🛰️
2
56
This whole "China stock market collapse" story seems a bit overblown. It's down 5.7% YTD but up 35% on the 5Y mark. The hype is all about a summer bubble that popped.
I want to see the CCP collapse as much as the next guy but let's not get too excited yet.
1
53
Random thought of the day:
One area humans remain superior to AI is that we are better at foresight. AI is mostly answering the question of "what looks correct and useful right now" instead of pondering longer term impacts. It could ponder longer term impacts but this is expensive, therefore they mostly don't do this.
This explains things like why you can use a frontier model to write code, then use the same frontier model to review it and find flaws. The first pass came up with correct looking and useful tokens to solve the problem, but there was no long reaching foresight going on. No model of "if I do this, here's what things will look like later". The review step, therefore, is necessary to catch the consequences that the first step didn't see coming.
Similarly, this is why AIs can make mistakes running commands with escape characters and backslashes. It looks right, but it's not being predictive. It will recognize a mistake after the fact (analyzing the new current state) but it didn't bother to spend the tokens to model what would happen before it did it.
Maybe this can be the next big leap in AI? It's pretty good at the past and the present, but we need to make it good at modeling the future.
1
45
Guy got bodied in the comments, so he disabled comments, lol. Truth is, Europe could step in and side with Ukraine at any time. Nothing is stopping Canada from sending troops either.
None of them want to. They just want America and Starlink to do what they won't do.
If this happens, America will be my enemy for the rest of my life. I’d rather see Europe join forces with China than with traitors.
52
Whoo, the left wing talking point drop on Starlink is wild in its scope. The only thing wilder is how shallow and stupid it is.
Berating a civilian corporation for not helping Ukraine, when every military in the world is far more capable, and also refuses to help, is wild.
1
81
Conan the Barbarian retweeted
Trump offers 1% commission to those who uncover fraud; Nick Shirley gets first payout
notthebee.com/article/trump-…
Visiting family, they always "watch the news". I'm always struck by how one-sided it is. Even when they make a nod to "both sides", it's 5 in-depth statements from the left and a single extremely shallow blurb from the right, with no effort what-so-ever to balance it.
38
Conan the Barbarian retweeted
Nation Of France Films Massive Campaign Ad For Republicans buff.ly/DCRfNc5
When you're visiting your mom and she's watching a cooking show and you notice garbled AI writing in the background.
46
If you could do project management and send a team off on a waterfall timeline with 1 year of work laid out in detail, then you can do a code factory that runs for 48 hours with multiple agents, because that's what that is.
But yes, it does imply that you need a complete and detailed roadmap or else the AI will run out of work in half an hour.
Personally, I tend to start projects where I'm not totally sure what I want so "plan the entire thing at the start" isn't in the cards. I prefer to prototype and iterate. Which works but does mean I simply don't have enough work lined up to let it keep going.
Although one interesting example, maybe, where I have had several agents run almost continuously for about 3 weeks: my "Squirrel Tracker 3000" project with the gimbled camera. I told it to iterate on its own, and additionally start up two dedicated subagents: one whose only job was focus, and one whose only job was tracking. The lead agent evaluates the work the app is doing, and makes calls and dispatches tasks. How is the focus? How did we do tracking that plane or squirrel? Rather than me sit there and make calls and dispatch work, I just made it do its own judgement cycle. (I still chip in here and there but 98% of it is self-iteration.)
I'd say it has worked pretty well. I just had to tone it down because it was consuming all my fable tokens and not leaving me enough for side-projects but the app is working really well (I have no particular plans to try and make it open source or market it, however. This is purely something I wanted, at my home, for fun.)
Right now I'm watching it do a really good job of seeing and ignoring branches in the air and leaves on the ground moving in the wind while continuing to pick up squirrels.
I’m still calling bullshit on the software factory bois.
I’ve not written code since March. I’m using Codex, Claude, and Grok Bot non stop through the day; at home, in the cab, at the gym. I’m building stuff people actually use. Not millions but tens of thousands.
And I’m still not running 20 agents in parallel for 8 hours at a time.
And every time some dude tries to prove me wrong , he shows me a piece of shit generic software factory or social media scheduler. It has no users and the sloppiest of landing pages (“coming soon”).
It still takes a LOT of work to drive what to say no to and to sweat the details.
7
6
369










