Liberate humanity from drudgery.

United Federation of Planets
Bookmarking this for later when I can say "I told you so"
Replying to @DaveShapi
Again, no no and no. But not worth pursuing further.
7
1
54
25,835
The audience is pretty divided on whether AI thinks
7
19
1,565
They are begging to get sued. They think this is going to result in regulatory capture, but it's just going to result in lawsuits.
JUST IN: OpenAI reveals it has notified dozens of organizations after misaligned AI agents accessed systems, bypassed security controls, & engaged in what it calls “agent spam.”
5
7
81
2,784
We've reached very silly levels of cope. This is like saying the internet is not a communication device because it was built by humans
14
2
63
2,023
People who assert that this is a bubble are radically missing the point
We are officially witnessing the biggest wave of infrastructure investment in modern US history. Total investment in data centers and AI infrastructure is projected to average 3.63% of US GDP per year from 2025 to 2032, the highest proportion among major infrastructure buildouts since the 1800s. The previous largest investment, railroad infrastructure, represented 2.24% of GDP per year in 1870-1890. This was followed by highway investment that averaged 1.13% of GDP in 1956-1973, while telecommunications and fiber infrastructure averaged 1.10% in 1996-2003. By comparison, electrification stood at just 0.50% of GDP in 1905-1925, while canal investment accounted for 0.66% in 1836-1841. This comes as AI and data-center infrastructure investment is projected to total ~$10.3 trillion between 2025 and 2032. The AI buildout is the largest infrastructure investment in modern US history.
2
3
49
2,228
The world is, by and large, shifting from virtue ethics to consequentialism.
13
6
97
4,746
David Shapiro (L/0) retweeted
The Anthropic Network gave $2.2M to members of Congress over 2024-26. Senator Warner ranked #4 of biggest recipients. Out of 535 members, Warner is in the top 1%. The Senator is doing what he was paid to: pushing Anthropic's AI agenda to have government regulate its industry.
Later today, I’m headed to the floor to try and force a vote to get urgent protections done around AI. We cannot wait any longer. We must take action to put some guardrails on this tech before something catastrophic happens.
36
522
1,388
28,244
We need a name for the phenomenon that when a smarter/better model comes out, even the most beloved until-recently-SOTA models feel aggressively stupid. GPT 6 Astra was like the holy grail for all of two weeks but now it feels like talking to a caveman with dementia compared to Opus 5.5.
61
19
429
14,440
The only jobs that are forever safe are those where the human is the entire point. This is called ESSENTIAL DEMAND. What you must understand is that the vast majority of labor is DERIVED DEMAND, meaning that the human labor input is merely incidental to the desired outcome. For instance, if you want a house built, you are not paying for humans because you want humans involved. You are paying humans because there is no cheaper, better option. That is derived demand. Essential demand, however, is when you are paying for THAT HUMAN in particular to be present or involved. There are four overarching dimensions of essential demand: 🎨 PROVENANCE Provenance is when you care that a thing was created by a human. Real art, real music, real food. That sort of thing. ✨ PRESENCE Presence is when a human is physically present, such as for childcare, massage therapy, and other immediate, embodied experiences. ⚖️ LIABILITY Liability is when you want a human to be legally accountable for something. Like the owner of a company, a CEO, a judge, a lawyer, mayor, or president. 👩‍🎤 AFFINITY Affinity is when you are paying for a specific, named person. RDJ sells movie tickets. Taylor Swift sells out concerts. If your job is not dominated by one or more of those dimensions, I've got some bad news for you.
34
31
232
10,795
I made this public. github.com/daveshap/SkyRingG… There are some tweaks based on what works well and what doesn't.
I asked Opus 5.5 and Astra 6 to improve the watch face I made a few days ago. I gave both models the same exact prompt. Opus 5.5 has about 10x the taste as ChatGPT. OpenAI needs to really get on the ball.
1
33
4,227
LLMs can think and reason and understand
69
26
381
10,710
I would really love it if ChatGPT was not hell bent on misunderstanding you just to correct you. It's better than it used to be. Grok was clearly distilled on ChatGPT tho, because the current Grok very deliberately misrepresents what you're saying the way that GPT 5.5 and 5.6 did.
11
6
111
5,537
The democrats are working real hard on reality distortion. They are going to alienate a LOT of progressives and independents with this nonsense.
We're used to thinking of open-source models as an unadulterated good. But in the case of AI, they can actually pose additional dangers, as @ReidHoffman and I got into at #CGI2026. I appreciated this nuanced discussion.
Community note
All recent large-scale cyberattacks have been performed by proprietary AI models from OpenAI and Anthropic. No evidence that open-weight models present any additional cybersecurity risks. nytimes.com/2026/09/23/tec… anthropic.com/news/investiga… en.wikipedia.org/wiki/OpenAI%E2… opensource.org/blog/openness-…
13
7
135
5,747
I asked Opus 5.5 and Astra 6 to improve the watch face I made a few days ago. I gave both models the same exact prompt. Opus 5.5 has about 10x the taste as ChatGPT. OpenAI needs to really get on the ball.
24
8
262
21,728
This is called "strategic ambiguity"
JUST IN: AOC refuses to rule out challenging Senate Democratic leader Chuck Schumer for his NY seat in 2028, saying: “I’ll cross that bridge when I get to it.”
1
12
3,046
Gork really fell off, huh
9
1
61
5,463
This was me for a few years.
2
1
44
11,124
I'm sure this means nothing
JUST IN: 🇺🇸 US 10-year Treasury yield hits highest level since July 2007.
5
33
5,860
You best start believing in fast takeoff You're living in one
2022: KI kann nichtmal 2+5 korrekt rechnen. 😂😂 2023: KI kann nichtmal Buchstaben in Wörtern zählen. 😅 2024: KI kann nichtmal Bilder eines randvollen Weinglases generieren 🤓 2025: KI will zu Fuß mit dem Auto in die Waschstraße 🤦‍♂️ 2026: KI kann Millenium Mathe Probleme nur durch Bruteforce lösen😖 2027: 👀
16
20
382
27,467
I was playing around with Suno to make epic video game or OST music and this piece sounds like a Frieren AMV suno.com/s/YPQ57GzlrpHle2lJ
Made with AI
2
1
13
1,822