podgorniy
Soviet doctrine treated Marxism as an objective, scientific law (I believe that's partially how the ruling class enforced to legitimize its power). Combined with post-WWII competition in space and computing, this mindset institutionalized rigorous STEM education, creating a massive pipeline of uncompromising textbooks, teachers, mathematicians, and engineers.
This partially explains why egineers from former-ussr places are good. They are taught by those and with materials which were produced by above mentioned processes.
That't the key. Not AI.
Who do you want to educate from that position? You or others?
Yeah. Like renting a house for a mother of 2 children who will never get her own by selling her labour, while the renter uses her money to buy more units.
Think again. The demand from rich to park wealth in X, Y, Z makes prices go up. Just imagine that the RAM prices fueled by AI spending is a housing market. More money rich have, less affordable the life for the rest is.
And don't forget that speed of their wealth growth is way faster than anyone who sells their labour.
Ah. And we did not even talk about "hidden" wealth thingy... All these numbers are official ones.
--
If person rationalist they
> prioritizes explicit reasoning, logic, and evidence over intuition, tradition, or emotion.
By sticking to this principle one has to describe every cause-effect, every idea from their speech. Reality is way more complex and detailed than any of it's representations. So rationalist can keep detailing and grounding everything they touch in their speech, and then ground grounds and so on. Sticking to rationalist values you have to write more and ground more.
By cutting corners of explanation and reason rationalist's claims they becoming less rationalist and less acceptable by the group, thus less popular. So the more you're rationalist, more popular you are more you have to write. It's natural way of things.
--
Irinically the complete rationalism isn't achievable because Gödel's incompleteness teorem.
> There will always be true statements about symbols that the system can write down, but cannot prove or disprove using its own rules.
They, we all have to draw a logical/grounding line somewhere, either by choice either by incapacity.
> we really want your data
this is a projection
They briefly articulated one obvious, undisputable and toothless truth about themselves. This is far, far from "extremely"
Paper describes such system whch has ability to enable/disable capabilities without process restart. Each plugin must follow specific shape: function to activate, function to deactivate it (both working, mutating the same shared context object), describe provided and required services. This, plus some ideas, like "every provider must outlive consumers when destructred" allows plug-n-play ensuring all dependencies are satisfied before the plugin is activated, lazy activation, and plugin deactivation/destruction.
All these discussions about AI to me like discussions about sticks. We need to talk about and deal with humans in the first place.
I think they did not have access to all pleora of enigma-related bits and pieces. Or there were not enough autistic ones. Or this one was simply overlooked in favour of more interesting one.
The whole trick is possible only because bunch of people whote bunch of text and code about the subject, well-documented it and made public. For LLM all these bits and pieces are very "close" and easy to pull together unlike for people who have to deal with each bit and decision and information.
I did another proper run and gemini 3.8 flash in antigravity solved it in about 45min https://news.ycombinator.com/item?id=49805363.
with search it found and referenced pages, including the hn ones. Without search it just described what one would need to descrypt (`To decrypt this ciphertext, the specific Enigma machine parameters are required:`) and the list.
Out of curiosity ran the same prompt against bunch of models - grok, kimi k3. They all say the same thing that they need model version, rotors and so on to descrypt.
When file output tool is enabled, some models give python script.
--
I read through some of the logs that antigravity gives. It produced intermediate results, scripts, calls, assumptions (about german language). I've shares random bits in comment below to give a taste of what it was doing.
--
The freshness of the news reduces changes that model fetched response from them
--
it searched for enigma-related repos and implementations, fetched various github repos parts, build inline descryption program.
It ran bunch of various scrips like:
clang++ -g -fsanitize=address /Users/dp/.gemini/antigravity/brain/e9a54e5f-1325-448a-8d43-fc537b901f34/scratch/enigma.cc -o /Users/dp/.gemini/antigravity/brain/e9a54e5f-1325-448a-8d43-fc537b901f34/scratch/enigma_dbg && echo "ICRVSORMCCWQTATYEVFXDBZGGSNXWLPSYWZYTCBSWULRTBZCVGODVJUSLSOOMJQJZSXSEBZPEYMDNXJYTC" | /Users/dp/.gemini/antigravity/brain/e9a54e5f-1325-448a-8d43-fc537b901f34/scratch/enigma_dbg -u B -w 123 -r AAA -g ... -c -l /Users/dp/.gemini/antigravity/brain/e9a54e5f-1325-448a-8d43-fc537b901f34/scratch/english
and
sed -n '1060,1130p' /Users/dp/.gemini/antigravity/brain/e9a54e5f-1325-448a-8d43-fc537b901f34/scratch/enigma.cc
and
Running 82M combination scan for unsteckered Enigma across all rotors, reflectors, positions, and ring settings. Monitoring progress.
and
Scanning all 60 rotor permutations and reflectors B and C across all ring settings (step 2) and all 17,576 indicator positions. Monitoring progress.
--
I also have opus running. It produced some sypher cracker which is still running (estimated time 100min, is about 15 min left)
--
My point is that astra isn't special. This appears to be quite narrow, well-documented and explored task. The goal itself is approacheable by other LLMs and non-researches task.