Humanity's Last Programmer; Game Designer; Problem Solver; past: OpenAI (Dota), Pro Competitive Programmer, Poker
Joined May 2012
- Tweets 2,215
- Following 393
- Followers 27,497
- Likes 1,639
352 Photos and videos
Pinned Tweet
16 Jul 2025
Humanity has prevailed (for now!)
I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive.
I'll post more about the contest when I get some rest.
(To be clear, those are provisional results, but my lead should be big enough)
549
1,098
13,140
2,219,134
16h
I've decided to stream my blind playthrough of Order of the Sinking Star demo (new Jonathan Blow game) on Tuesday, June 16
What's the best time to start the livestream?
it will happen on twitch btw: twitch.tv/fakepsyho
0%
UTC 16:00 CET 18 EDT 12
0%
UTC 17:00 CET 19 EDT 13
100%
UTC 18:00 CET 20 EDT 14
5 votes • 7 hours
6
1,161
20h
Are there any platforms where we can bet that Fable 5 is going to be back before stock market opens on Monday?
At this point, I'd give it like 30-50%
10
2
71
10,258
20h
If you're sad that Fable 5 is no longer accessible, I'm happy to report that Fable 1, Fable 2 & Fable 3 are still available via Game Pass
4
5
154
5,315
Jun 13
This is the craziest part btw - it could pretty much paralyze the whole company.
When I worked at OpenAI, the share of non-US nationals among researchers was 40–60%. I would expect that number to be very similar today.
28
9
246
16,464
Jun 13
It honestly might be the best gift that the US government gave to Europe (and China) in a really long time.
I also expect this to rollback very soon. Access to US nationals only makes no sense. It should be either everyone or selected governments/institutions.
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
16
3
98
13,487
Jun 10
Random idea: I could livestream (twitch?) my blind playthrough of the demo while talking about puzzle design.
Cons: My English is terrible these days, so it might not be very entertaining
Pros: You’ll feel smart watching me get stuck on a very easy puzzle
Yay or Nay?
Jun 10
The SURPRISE is:
x.com/SinkingStarGame/status…
Order of the Sinking Star will have a free demo on Steam, for NextFest, Monday.
Because the game is HUGE, the demo is huge: it's bigger than most entire paid puzzle games, and you get to try it out for free.
5
1
35
3,512
Jun 7
Happy anniversary to the one of the most important AI papers from 2025
12
11
183
23,422
Jun 7
To be clear, I really do mean this unironically.
It’s rare for a paper to be widely shared, make bold claims, and still be almost trivially identifiable as complete bullshit.
Signal boosting this paper was like outing yourself as a disinformation spreader.
2
2
86
9,881
Jun 4
Opening this blog post with a LOC graph makes me assume the only goal here was to boost your IPO valuation.
I just cannot fathom how any engineer looked at this and thought "yeah this looks reasonable"
Jun 4
Today, Anthropic engineers on average ship 8x as much code per quarter as they did compared to 2021-2025.
15
5
146
10,406
May 28
I've been learning English my whole life, and I still refuse to accept that "there", "their" and "they're" are exact homophones.
5
1
34
5,414
May 28
This post was not sponsored by BoldVoice. But I do highly recommend it if you're working on your accent.
1
7
1,520
May 27
Crazy is an understatement. ~10 years ago, I had my own small indie studio and we were working on a game in Unity 4.x.
If I allocated any temp object in the game loop, I had a random GC call every ~10s that took 10-15ms, which was enough to skip a whole frame. Completely unacceptable in any platfomer, beat 'em up, etc.
You couldn't use strings (despite Unity API itself requiring strings as arguments), couldn't resize any dynamic data structures. Couldn't even use a foreach statement and you had to do memory pooling for the tiniest objects. Big chunk of the language was not accessible to you and AFAIK there was no workaround.
Writing your own game engine wasn't such a bad idea at the time.
May 26
Using a garbage collector in a videogame is crazy.
It’s a miracle unity succeeded by putting a gc into every produced game.
16
11
447
62,969
May 27
(apparently) unpopular opinion: if you're using claude code / openai codex, you're not coding, you're managing AI coders
19
72
9,088
May 27
if you have negative feelings about this statement, you probably think that coders are better than managers
2
23
1,781
May 27
if people really get problems like this during code interviews, I feel that the hate towards leetcode is fully deserved.
If you know the segtree (or similar) this is trivial. If you don't, you essentially have to reinvent the wheel.
24
3
238
40,626
May 27
A bit unrelated, but I honestly don't think I've used segtree in even a single heuristic contest.
2
20
5,903
May 26
I'll never understand why leetcode got so popular.
It's like competitive programming without the fun part. All of these are your standard textbook problems. Either you have the relevant knowledge and it's trivial or you can't do much. No problem solving required.
44
10
447
70,046
May 19
I guess Karpathy's timelines got much shorter
May 19
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
2
1
145
18,122
May 14
FYI, all four subgames are heavily inspired by existing free puzzle games:
- The Hearty Heroes of Hauling: sites.math.washington.edu/~o… sites.math.washington.edu/~o… sites.math.washington.edu/~o…
- The Mirror Isles: alan.draknek.org/games/puzzl…
- The Promise: silverspaceship.com/promesst… & silverspaceship.com/promesst…
- Skipping Stones to Lonely Homes: alan.draknek.org/games/puzzl…
Heroes of Sokoban is my favorite and also the most approachable
3
10
307
36,844
May 13
protip: if you see AI and IQ in the same sentence, you can be sure it's just pure webslop cosplaying as science
May 12
Today I’m launching AI IQ — frontier AI models, scored on the human IQ scale.
Instead of endless leaderboard tables, AI IQ shows:
• Where models land on the IQ bell curve
• How frontier IQ is changing over time
• How models compare on IQ and EQ
• What intelligence costs in practice
GPT-5.5, Claude Opus 4.7, Gemini 3.1, Grok 4.3, Kimi K2.6, Qwen3.6, DeepSeek V4, Muse Spark, and more.
Link in the first reply. Curious which chart surprises you most.
7
143
11,219