E_Bfx

Yes but it only one face to face meeting is needed in the process to see if someone is using AI to answer interview. The 13 other interviews can then be online.

That was my first thought : "Oh NFT is still a thing in 2025". But no, it is really awesome.

That's the problem, if you are working in a compagny which have mostly junior (1 or two year of programming), it is better for you to not implement to complicate pattern otherwise your day will be fill of explaining what a Factory is.

Western country are self suplient of food because we have engine and chemical industry, ones this goes down, shortage of food will be quick to come. In part of the world where labour is still mostly manual, it will be more resilient.

Also without bank, a lot of people will find themself without any properties and will likely get more violent. Again, in part of the world where people own real objects and not number in a computer in a datacenter, this won't happend.

Switzerland has no oil and depend on its banking sector. If USA crashes then likely all modern supply line will be cut and finance will be something of the past. The current well being of a country does not indicate much how it will survive a global crisis. A farmer in rural Africa would be less affect by the implosion of USA than a trader at Geneve.

Also conda does not have a free licence, every organisation with more than 200 employes should pay for it (https://www.datacamp.com/blog/navigating-anaconda-licensing). uv is at date, Apache 2.0.

I guess that there is something with the cache that we don't do right. Thanks for your return.

In a medium size codebase (~ 100 python modules of 200 lines), mypy take 5 minutes to type check. This can be a problem for a CI.

In my organisation, some co-workers used to write def func(*args,**kwargs) all the time. That was so tiring to look for what you should put as argument. Type checking is mandatory for well organized large project.*

- There is also (but on a smaller scale) a gamification of math with bounties (https://mathoverflow.net/questions/66084/open-problems-with-...) but when a result is proved you cannot prove it "better than the first time". So it is more a "winner take it all" situation. - I am not sure but the "2-minute papers" equivalent would be poster sessions, a must-do for every Ph.D. student - For the marketing side, there are some trends in math, and subtly researchers try to brand their results so they become active research fields. But since it cannot be measured with GitHub stars or Hugging Face downloads, it is more discreet

Your website is amazing, great job

> Transformer² represents a significant milestone in the evolution of AI systems.

Coming from a math background, it always amazes me to see how people in AI/ML brag about their papers. If someone wrote:

> My paper represents a significant milestone in the evolution of algebraic geometry/ergodic theory/combinatorics

it would be a laughing stock for the math community.

pymupdf has a commercial licence that couldb be a problem if use in a compagny.

Very impressive, the Python ecosystem is slowly getting very good.

I guess it is tautological from the definition of "provable". A theorem is provable by definition if there is a finite well-formulated formula that has the theorem as consequence (https://en.wikipedia.org/wiki/Theorem paragraph theorem in logic)

Saying that working with AI is just "LLM prompts and orchestration" is the same as saying that coding is no more than some for loop with sometimes if condition. Technically true but irrelevant.

Thanks a lot,

Most people think it is easier to use EC2 than FarGate since the first is the most famous one. But actually, it is the other way around!

> drunken socialist sailors

Sorry, English is not my first langage what is this expression ? Why does the sailor as to be socialist ? Google didn't help me with this one.

Hello, could you elaborate on the footgun with mixing sync and asyncio ? I am currently devellopin one app with both kind of endpoint + websocket and I prefer not discover too late this problems .

How easy is it to switch from OpenAI to testing a LLM on premise ?

'plan' is an ambigous word. 'plan' for Devin means break a complexe task in subtask and then generate code. Whereas 'plan' in the paper you link is never define, so I am not sure what his author want to demonstrate, that LLM doesn't have free will ? that LLM are not universal problem solve ? It is a quite confuse paper.