Prdctnomics - The economics of technology products
@prdctnomics
Followers
2K
Following
65K
Media
272
Statuses
3K
Building LLM products by day, studying the formula that creates great companies by night. Product, Software, Technology Mega Trends. Not investment advice.
0 to 1
Joined November 2010
Four years ago, humanoid robots were laughable research toys that fell over during demos. Today, they're attracting billions in investment dollars and sparking a technological gold rush. Why now? Three forces have collided: 1. AI finally works. The same breakthroughs powering
1
1
3
Excellent foresight as usual from Jon Stokes.
I'll go ahead and put a flag in the ground here. This AI 2027 thing is wrong, the intuitions behind it are bad, and in 2027 these guys will publish a post-mortem that mostly amounts to "we were right, just got the dates wrong".
0
0
0
Claude Code was a side project at Anthropic. ChatGPT was a side project at OpenAI. PyTorch was a side project at Meta. Gmail was a side project at Google. Side projects are the only place where taste, curiosity, and agency fully compound.
180
628
7K
There is finally so much ai product coming out of google that it is hard to keep up.
🚨 BREAKING: Google Research just dropped the textbook killer. Its called "Learn Your Way" and it uses LearnLM to transform any PDF into 5 personalized learning formats. Students using it scored 78% vs 67% on retention tests. The education revolution is here.
0
0
1
2026 is going to be wild.
Despite its high 50%-time horizon, Opus 4.5's 80%-time horizon is only 27 minutes, similar to past models and below GPT-5.1-Codex-Max's 32 mins. The gap between its 50%- and 80%- horizons reflects a flatter logistic success curve, as Opus differentially succeeds on longer tasks.
34
65
947
The growth of Google Gemini is insane.
ChatGPT by OpenAI was the fastest app ever to hit 1 billion users, but Gemini by Google is now on track to smash this record. Should OpenAI be worried?
41
82
980
We estimate that, on our tasks, Claude Opus 4.5 has a 50%-time horizon of around 4 hrs 49 mins (95% confidence interval of 1 hr 49 mins to 20 hrs 25 mins). While we're still working through evaluations for other recent models, this is our highest published time horizon to date.
67
278
2K
Gemini 3 flash is as good at reading handwriting as the average human (pro is expert human level). It is much better than both GPT-5.2 and Opus 4.5 with character level error rates of 1.43% and word level error rates of 2.74%. This is a 47-63% improvement over 2.5 Flash, the
27
103
975
I encourage folks to watch this video and then realize that's just the baseline expectation/cost for a Chinese EV (with many more higher end or lower end options). They completely blow out most American car companies of the water on quality and cost. China made an intentional
I lived with a Chinese EV for a few weeks to see if the hype is real. The car costs $42K, and turns out it feels like $75K+ EASILY One of the most impressive things I've ever reviewed: https://t.co/EHe358cvLt - A+ software and features. Feels like what would happen if Apple
20
99
919
Claude code rivaling OpenAI in it's revenue while not having anywhere near the level of customer adoption is again confirmation that enterprise always wins at the end of the day. If you nail the enterprise use case down with repeatability, the world is your oyster!
21
7
282
Huge opportunity here
companies like Facebook record every imaginable interaction their users have with the platform. they log each of your clicks and taps. they keep track of how long your gaze lingered on a post, whether you were on the same WiFi as that woman who might be your friend, which
0
0
0
A few graphs that show why Google would be smart to "burn the boats" and go all in on Gemini: 1. Google share of search sessions is finally starting to fall to ChatGPT and other LLMs - now <80% vs. 88% this time last year. Google's DAU/MAU ratio has also moved downwards for the
33
43
339
We’ve been intensely cooking Gemini 3 for a while now, and we’re so excited and proud to share the results with you all. Of course it tops the leaderboards, including @arena, HLE, GPQA etc, but beyond the benchmarks it’s been by far my favourite model to use for its style and
219
501
6K
I thought Satya's answer on why they paused datacenter buildouts in 2024 was potentially the most interesting part of the podcast "We're not in the business of just doing 5 contracts with 5 customers being their bare-metal service. That's a business for someone else" Ice cold
38
59
887
This is actually the most damning part. Zuck is being played by people who have no faith in him or his company. Psychopaths are generally undesirable, but if the CEO isn't one while his underlings are… ngmi
Butterfly effect: one of the last convos I had with Mark in the beginning of 2023 was pushing to create a smaller FAIR v2 in order to ruthlessly pursue AGI (while YLC was in the middle of rebranding it to "mon AMI" for "Artificial Machine Intelligence"). Mark's response was that
10
16
403
'The securities exchange model is so powerful that nearly every exchange with a 20-year public track record has outperformed its respective regional stock index, in most cases by a wide margin. It’s a global phenomenon: The U.S., Japan, Hong Kong, the UK, Singapore.' - Horizon
Fresh off the presses, our latest Quarterly Commentary offers a deep dive into everything from localized inflation plays to limiting-factor beneficiaries of continued AI/data center growth: https://t.co/6yKt1d23PH
11
24
232
Easily one the most entertaining drawdowns of my career $CSU.to
13
9
160
One of the biggest mistakes is that folks lump everything into "AI" There are at least three important and distnct buckets right now; 1) big horizontal models, 2) apps built on these models, 3) actual vertical AIs with their own training data within a specific domain -- drug
science.org
Fully automated system is working 24/7 to create new inorganic materials that could improve batteries, fuel cells, and superconductors
The two biggest narratives about AI right now are that (1) it’s a massive bubble (i.e. totally fake) and (2) about to give rise to superintelligence (i.e. totally real). Both narratives can be fake (which is what I believe) but it’s very unlikely that both can be true.
2
4
15
Warren Buffett has said goodbye. His final words as CEO of Berkshire Hathaway:
4
21
60
Indeed. World getting closer to accepting most obvious econ implications of AI: things it can do will get very cheap, complements to those will get expensive, society will get richer, things with + income elasticity of demand will get more expensive. Scarce factors get rents. 1/2
How the Baumol effect and Jevons paradox are related & what they mean for the AI supercycle By @Alex_Danco Full piece in the a16z newsletter
7
47
335