Nico Szerman (@nicoszerman) profile picture

Founder @sailingprotocol. Ex Engineer at Meta. I design and build financial infrastructure. 🇦🇷 🇺🇸

Boston, MA
Joined January 2012
Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)Photo by Nico Szerman (@nicoszerman)
Less HODL more BUIDL
4
1
37
Nico Szerman retweeted
Pretty much all theories that predict huge maybe-world-ending events also predict many smaller related events, usually in a power-law distribution. So don't tell me your big theories don't have testable consequences short of the huge event you fear.
45
34
21
386
56,063
It’s too soon to observe transformative AI in the data. Wait two years.
I've gone back and forth on how transformative AI will be. But the more I look at the underlying economic data, the more I'm convinced we're not getting either utopia or doom any time soon. Consider: -GDP growth is normal, and projected to remain normal. -The stock market looks pretty normal. It's done well in recent years, but there's no indication that we're getting massive increases in societal wealth due to rapid productivity gains. -Economic theory says transformative AI should massively increase interest rates. If massive growth is coming soon, people will want to borrow against future wealth. If we're all going to die, no one will lend. In both cases, interest rates go up. Yet interest rates look... completely historically normal. -There is nothing in the jobs data to indicate widespread displacement of workers, not even a major shift in the workforce, much less "the end of work." Practically all data is pointing in the same direction. AI looks like a normal technology. I think that those who said that the bottleneck to changing the world wasn't book-type intelligence are looking like they were correct. We keep hearing that AI bulls have gotten so much right, but although they've been correct on the pace of technological development, we're yet to see the real-world impacts they promised. It's time to start thinking carefully about why that is. richardhanania.com/p/ai-is-l…
56
Nico Szerman retweeted
We need a CLEAR positive narrative for what humans will do post AGI. We have indefinite optimism which is better than doomerism But we need DEFINITE optimism and a new clear view of humans post AGI. Urgently.
People should always be in control of the future. OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that “AI systems think makes sense to reward in humans”. People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress.
11
7
2
57
5,364
Nico Szerman retweeted
As part of my campaign to point out that discussions on defensibility are generally pointless, here is a very classic theoretical result showing that you can't even efficiently compute Nash equilibrium states.
19
9
3
140
16,811
No significant difference in growth rates pre AI vs post AI, wow!
At @Stripe Economics, we looked at industry-level US productivity growth. There is indeed a weak relationship b/t AI adoption & growth since ChatGPT (top panel) But AI adoption also predicts *pre-pandemic* growth (middle) Diff has no relationship (bottom) stripeeconomics.com/p/ai-and…
1
2
97
This is certainly temporal until models get better
1
5
Nico Szerman retweeted
You can now book our humanoid cleaners on a recurring subscription, for up to 3 hours per cleaning at tau-robotics.com. Video is first take, uncut, shown at 4× speed. See the reply for detailed footage at 1× speed.
929
789
735
8,606
2,072,198
So many people building shit products and complaining they can’t grow because they lack distribution.
2
41
If you see software as merely "crystallized intelligence," as @CliffordSosin said today, then similarly to how senior professionals are much more valuable than superintelligent 1 year old babies, advanced software is much more valuable than general superintelligence.
1
43
Specially given that the advanced software will be made by superintelligence, so the comparison is a superintelligent 40-year-old vs a superintelligent 1-year-old.
12
Nico Szerman retweeted
What is life like at different levels of GDP per capita? In this new chart I visualised it for 12 different dimensions.
80
628
130
2,488
262,149
Nico Szerman retweeted
Donald Knuth on P=NP:
31
58
11
1,246
129,790
yes i think so: like all knowledge domains, models saturate that domain over time (this already happened for translation 2 years ago, for example). i think this will happen in cyber roughly the middle of 2027 at the very high end of model intelligence
7
9
4
251
42,298
We got super-dog dogs before super-human humans
SITUATION DETECTED: Warhound launched from stealth today with a $200M seed at a $1.2B valuation. Ben Lamm and Palmer Luckey are cloning gene-edited bio-enhanced dogs and equipping them with custom Anduril vests, goggles, and sensors for the military and law enforcement.
1
41
the new a16z.com website looks SICK
1
41
How is AI psychosis not the most studied condition in psychiatry today. It may be a new mental illness! andys.blog/starting-over
1
34
Steve Jobs emails Paul Rand March 2, 1996
48
107
52
4,717
974,003
Abundance LFG
Abundance usually gets explained as a list of problems: zoning, permits, licensing, slow approvals. Here’s what we'd actually get if we fixed them: Lunch costs $4, like it does in Japan and Singapore. Most US cities require three or four separate sinks, which makes a kitchen under 200 square feet nearly impossible. Anyone can open an owl cafe or a vegan lesbian bookstore. Japan is full of tiny, oddly specific businesses run by one or two people. They survive because small spaces are cheap and enough people live close enough to walk in. A room in a big city for $400 a month. These buildings housed millions of low-income and older Americans until cities banned most of them by the 1980s. A three-bedroom apartment you can stay in after you have a kid. US codes require two staircases in buildings over three stories, so family-sized units only fit on the corners. Most of the world allows one staircase. A new electric car for about $10,000. China sells these globally. Americans can't buy it. A government so dependable it's boring. When a loved one dies, you report it once instead of to every agency. We can decide to make it happen here. The full list has 16 things, from taco trucks on every corner to no more pandemics. I wrote this to also announce that today is DC Abundance's official launch! If you’re in DC, come celebrate and see the G.O.A.T. Derek Thompson speak!
39