0xidative retweeted
3 reasons:
1. AI slower, less corp bonds, less supply on the long end
2. AI slower, US GDP slower, real yield down.
3. AI slower, less memory/computer component supply shock, less inflation.
All points to lower US10Y
1. AI slower, less corp bonds, less supply on the long end
2. AI slower, US GDP slower, real yield down.
3. AI slower, less memory/computer component supply shock, less inflation.
All points to lower US10Y
@labubu_trader why? rates dropping fast after initial jump
15
17
231
62,466
Wikipedia golf benchmark makes a pretty intelligence test, I should make it
Game: race from one Wikipedia page to another using only links
Challenge: choosing between hundreds to thousands of links
Shows not just intelligence-per-second, but also the compounding benefits of not hallucinating with high-cardinality choices
0
0
1
49
My signal is doomscrolling instagram reels and when I see a ticker, eg $RDDT being posted frequently as a steep discount stock, it leans value trap rather than undervalued quality asset. These are all good reasons to avoid it
Not opposed to $RDDT at this price but this is one I don’t like spending time on because:
Reason 1: Investor base too immature
I told you when I bought Reddit at the low this year of $124, cost basis optimized some higher, and on the posts about selling, got so much hate. Including personal attacks.
I guarantee you if the same was done for say, $NVR or $SAIA, you wouldn’t do that. You would not interpret any negativities as being about “your” stock and feeling the need to attack me personally, as anonymous trolls.
As such, this is a ticker I just don’t want to waste time talking about. Not worth it. I did add a little more yesterday albeit unenthusiastically.
Reason 2: Not going to write it out again, see quoted post.
0
0
1
101
Long $GOOGL
- AI opportunities across their whole org
- Youtube is a compounding machine
- Youtube music is undervalued
- Top 3 in terms of AI distribution opportunity
- Sundar is a pretty good CEO all things considered
- AI opportunities across their whole org
- Youtube is a compounding machine
- Youtube music is undervalued
- Top 3 in terms of AI distribution opportunity
- Sundar is a pretty good CEO all things considered
@kevinxu Google.
- Full-stack AI w/ custom silicon, DCs, fronteir lab, and distribution across billions of daily surfaces.
- Huge cloud backlog of over $500B w/ cloud biz growing 80%+ in Q2.
- Other fun stuff like Youtube growing, Waymo, Deepmind, SpaceX position, Anthropic position.
0
0
0
41
0xidative retweeted
@bee_fumo fair question. It's basically Minecraft's engine rebuilt from scratch in C and CUDA. All the same physics, world gen, and pixels, meaning I verified frame by frame against the real Java implementation, of course.
The reason is so I can run thousands of copies of the game at once on one GPU to train tiny neural networks. Real Minecraft is nowhere near fast enough, and this just re-implements it with thousands of worlds in parallel at many aggregate ticks per second, enabling models to learn and train on this very fast.
It also acts as a rendering farm if you use the not-so-cut-down version of the game and instead just have a much faster batch renderer for getting the same pixels that the Oracle would give you.
The reason is so I can run thousands of copies of the game at once on one GPU to train tiny neural networks. Real Minecraft is nowhere near fast enough, and this just re-implements it with thousands of worlds in parallel at many aggregate ticks per second, enabling models to learn and train on this very fast.
It also acts as a rendering farm if you use the not-so-cut-down version of the game and instead just have a much faster batch renderer for getting the same pixels that the Oracle would give you.
5
1
142
7,104
Update from Anthropic regarding outage:
Identified - We have identified an issue resulting in elevated errors across multiple models, and are working to resolve this. We will provide an update as soon as possible.
Jul 29, 2026 - 20:33 UTC
Identified - We have identified an issue resulting in elevated errors across multiple models, and are working to resolve this. We will provide an update as soon as possible.
Jul 29, 2026 - 20:33 UTC
0
0
2
162
Great message more people should hear
1/ Defense Reformation Thesis #1: Monopsony Is the Root of What Ails Us
Brilliant explanation by @smartereveryday of how monopsony crowds out commercial manufacturing.
When the government is the only buyer, things will not work, they will be expensive, and they will make us weaker.
0
0
1
21
$ASML cooked on earnings, but most important point is strong guidance through 2028 from management
0
0
0
38