thieme

Enterprise Architect. Making software teams thrive (AI, CD, Clean Arch, ATDD, TBD, agile, Rust, Haskell, ...), ML, scientific skepticism, meditation, WFPB diet.

[thieme @thieme]

@burkeholland There is more than reading/not reading. It’s not binary. Reading code can be 8h a day or 20 mins per week for just the critical parts.

[thieme @thieme]

@izzy2live @NXT4EU We can play this game and walk back further and further. It’s mostly emigrants from Europe who made inventions. Let’s go back to the printing press, which was before Europeans discovered their new land.

Many greetings 👋🏻

[thieme @thieme]

@unclebobmartin @kaio_m1 Interestingly now the current US gov is taking socialistic/communistic steps:
https://t.co/8qyxTnWSua

[thieme @thieme]

@MatthewBerman @endingwithali You are not an experienced dev, Matt, so for you it might not matter. But “reading code” is nothing binary that you either do 8h a day or never at all. We can still spend minutes to hours every week to check the critical parts.

The *endgame* will be: not reading the code.

[thieme @thieme]

@izzy2live @NXT4EU All those depend on Europe’s ASML, and Zeiss technology from Germany.

[thieme @thieme]

@izzy2live @NXT4EU Ukraine has the second strongest army in the world. They are the most cost-efficient army in the world. Not even the US has this practical experience and the technology that Ukraine has.

[thieme @thieme]

Europe should start its own crypto currency and its own companies, led by AI. Everybody will get the same amount of coins per month, for free, and can buy from those AI-run companies with those coins.

AI will soon take most jobs. This should be the new system. Fair for all.

[thieme @thieme]

@unclebobmartin @valigo At least for parts of the system this may already work.
It’s not frontier math... most should be fairly easily solvable.

[thieme @thieme]

@eddib17 @OfficialLoganK @edandersen This: yes. Seeing how far ahead they are. But in this case it wasn't about the "clear lead" by Gemini.

[thieme @thieme]

@MatthewBerman @BrianNorgard No, it's bad. Let the fish live. Or go find some that have already died, but don't kill other entities when it's not required.

[thieme @thieme]

@OfficialLoganK @edandersen How often are OpenAI and Anthropic using Gemini though to port games for fun?

[thieme @thieme]

@Sc_Meerkat @unclebobmartin @valigo Same as with AI-maintained code bases, right?

[thieme @thieme]

@LGrapenthin @unclebobmartin @valigo In your brain there is nothing but neurons. There is no classical software running. All your software is fundamentally made out of probabilistic computations.

[thieme @thieme]

@valigo @unclebobmartin AI labs have the time and compute to start a search for better code and algorithms and can train on that. Soon, I dunno, we might not need full test suites anymore and instead have an AI use Alloy, TLA+ and Lean and *prove* that our code is correct.

[thieme @thieme]

@valigo @unclebobmartin But they are trained on more. We have synthetic data with lots of defects and the correct fixes. In a code base GPT-5.5 immediately finds bugs that were missed by humans, for years, and in an env with thousands of tests and 4-eye-principle.

[thieme @thieme]

@gambarian @seanboisselle @ilyasut @grok When? And on what basis? You mean after Europeans lived in the US and made those inventions after they have left Europe?

[thieme @thieme]

@seanboisselle @ilyasut @grok please fact-check.

[thieme @thieme]

@michaelshermer Interestingly, @michaelshermer, we already *are* multidimensional energy beings.

This is because 3d is "multi" and mass equals energy.

[thieme @thieme]

@citadel05ricks @DeryaTR_ Most billionaires risked a lot and gave great inputs into the market. They now have a totally safe life.
In future ASI will do everything. No Great Ape (i.e. human) will be competent enough. So the AI should get everything and humans nothing? Let’s do a fair solution.