Solving AI-generated games that don't hallucinate. NativeAI.Games, & Ex-EA | @Multiverse_FTL | @StarChartApp | The Copenhagen Post.

Joined March 2009
511 Photos and videos
My guess is Howard Lutnick asked Fable to guess his password and it one shot it.
1
9
The password was probably "MAGA2028!"... right?
2
For our next product I've been working on a puzzle generator. *NOT* AI based and entirely dynamically generated. The goal is to reverse engineer how a puzzle game works so it can be applied to any game, and then skinned by hand or by AI.
4
I get it now. This is a ploy to get Dario ousted before the IPO. It all makes sense now.
11
HIRING: US citizen in the US to copy / paste my prompts into Claude Fable 5. Where there is a will, there is a way.
48
And so it begins... Sonet 4.6 just downgraded me to Haiku 4.5 mid conversation. First time ever.
17
This is why I never took an IQ test... The government would have forbidden me from talking to foreigners or leaving the country on holiday. 🤪
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…
18
Today's learning: AI needs you to have a very clear architectural vision and a goal.
1
2
35
Today's learning: There is a level of complexity where AI just gives up and melts into slop. It's not that much complexity either. And it doesn't matter if you write the code yourself, the first change needed messes up everything. I think it's knockons. It doesn't understand em
1
30
Today's lesson: AI as currently built is a civilisational-scale replication of human mediocrity, without any of the human qualities that occasionally overcome it.
1
2
84
In game dev nothing is truer than "Good enough is the enemy of great". Yet AI code is exactly and precisely "good enough".
2
73
Today's lesson: don't believe AI when it tells you it has done something. Get it to prove it.
66
The more I vibe code, the more I see both how great AI is, and how essential humans still are.
1
64
Yo @Grok. Just been reading up on X.com's embedding of users for throttling... I f**king hope I'm a vector shooting off at abhorrent right angles from the vectorised norm. Vectorise this 🖕
1
68
Vectorising people and then building vector cones that encompass a specific vector space is the new business model. In the end, we are just numbers to the man.
42
The logic: When you look at it... Trump isn't very fit. The Rock is fitter... So he should take all Trump's money. Etc.
May 15
Trump on Taiwan: When you look at the odds, China is very, very powerful, big country. That's a very small island. Think of it, it's 59 miles away. We're 9500 miles away. That's a little bit of a difficult problem. Taiwan was developed because we had presidents that didn't know what the hell they were doing. They stole our chip industry.
37
Today's learning: Vibe coding is harder than coding yourself. Things get done quicker, but they are wrong in the worst possible ways.
23
AI model improvements have visible gaps. For example one theory of why the GPT 4 series was so popular is that it was trained on copyright infringed books. While the GPT 5 series isn't. The result is intangible holes in capabilities the majority of the tech industry don't notice.
35