Mentioning neural ODE doesn't make sense here, as this is unrelated. Basically any implementation of transformer uses residuals, but you're not really training a neural ODE here.
Also consider getting rid of the em-dashes. I don't know if you mostly vibe-coded this or not, but the README is pretty clearly AI generated.
Good, coding harnesses should be open source and LLMs should be treated as commodities. Minimize switching costs for consumers, and let people understand how they're interacting with the context and the LLM outputs.
The industry has been moving the wrong direction with Claude Code staying closed (despite multiple times leaking the source code!) and the open source Gemini CLI being deprecated in favor of closed source Antigravity CLI.
is this the ChatGPT finance features they launched in May? it keeps asking me to integrate my finance data, but I have doubts about how useful it would actually be (not to mention some distrust about how well they would actually protect my data).
The interesting part is they chose to go with a normalizing flow approach, rather than the industry standard diffusion model approach. Not sure why they chose this direction as I haven’t read the paper yet.
As with many startups (especially ones with high burn rates) OpenAI is risky. It could take down SoftBank and its data center vendors. 6% of nvidia’s revenue is not that concerning, as I’m sure they can find other buyers for those GPUs. But I really don’t buy the argument that OpenAI is the gen AI industry. If they ceased to exist tomorrow, the tech/genAI industry would just trundle along. At this point the tech is quite commodotize.
I'd say it's more like kleenex. Lots of people ask you to 'pass them a kleenex' when their nose is runny, but they just mean tissue. They don't actually care what the brand is. Similarly for LLMs most people may not care (or maybe they will, and it will be more like Google search), especially if they just use it via some other app that calls LLM provider APIs. My anecdata so far says early adopters try multiple LLM providers and use the best one for their use-case. No clue on what non-tech folks think though.
Well this was an entertaining take on the job market.
tl;dr he's salty other people ended up better off doing the same work for other companies, based on views of tech compensation that are very divorced from reality:
> I look at my own place in the world compared to people who just started at Apple or Microsoft 20 years ago then never left, and now they have made eight figures just over the past 4 years while my life path has lead me to… practically nothing. Then the tech inequality continues to compound. Imagine joining a company where the teenage interns have already made a couple million off their passive stock grants and other employees have been making $2MM to $6MM per year over the past 5 years there, while you’re starting over with nothing again for the 5th company in a row so what’s the point in even trying
Nevermind that there aren't really any interns making a couple million off stock grants, and this part:
> Do we just sit here and die in our overpriced studio apartments where rent increases 7% every year while other ICs doing the same work at better companies are buying 5 vacation houses from doing the same work?
Also love the part where he implies he's too smart to pass coding interviews:
> According to all the interviews I’ve failed over the years (I don’t think I’ve ever passed an actual “coding interview” anywhere?), the entire goal of tech hiring is just finding people in the 100 to 115 midwit block then outright rejecting everybody else as too much of an unknown risk.
What ends up being useful is hard to predict, so it's better just to do what you enjoy. Lots of useful math started out as just an idle curiosity, though mostly it ends up being useless. Probably most engineering projects are the same though (most end up in the dustbin sooner or later).
Also consider getting rid of the em-dashes. I don't know if you mostly vibe-coded this or not, but the README is pretty clearly AI generated.