Hacker Newsnew | past | comments | ask | show | jobs | submit | wrenky's commentslogin

Totally anecdotal, but recently my wife had to go to urgent care for something wrong with her ankle- They send a 4-5 page sheet of arcane terms and diagnoses to her care app (relayed to me via text) and I just slammed that into gemnai and asked "what does this mean" and it did quite well! Gave possible causes, what it meant for her in the long term vs short term, and ways to prevent it. I had a better understanding of what was wrong before the doctor even got to my wife in the waiting room!

Obviously still double check things, but it was moment of clarity I hadn't really had before this. Still needed the doctor and all the experience to diagnose and fix things, but relaying that info back to me is something doctors are only okay at. Try it out! take a summary sheet of a recent visit or incident and feed it in.


The problem there is why they have such tight bottlenecks periodically- Why not just have a traditional cellar?


Great question!

Perhaps. Perhaps that was one of the use cases, but the inconsistent diameter could also be the result of them being bad at digging tunnels. Digging tunnels by hand is hard (and - dare I say - scary).


I don't think being bad at digging tunnels can explain the bottlenecks. They are way smaller than what you'd expect from inconsistent diameter due to poor digging.

Some inconsistency from poor digging can be ignored as a minor inconvenience. It is no big deal if in one section you have to duck a little. The bottlenecks on the other hand are so small that the noticeably impede progress, and some people probably could not get through them at all.

If they are not there intentionally that would be too big of an inconsistency to ignore.


They are an improvement over most languages, but the are _not_ better than sum types for the reasons listed above. Way less flexible and forces a lot more verbosity when compared to more functional languages.

Hes definitely wrong on how the error types can help account for that but it would be best in class if we could properly chain them AND use all the greatness from interfaces.


There is a good argument for never using debuggers except for core development- Once finished your logs/metrics/events should be good enough to understand what is happening in an application. If debugging your application requires breakpoints you wont really be able to debug a live instance, and wont be able to easily signal off what is happening in the future.


That is a reasonable argument - but it was not made in the article and also does not preclude the use of breakpoints (see your except clause which covers a lot of ground).


> Our hope is that these extensions can over time be contributed to upstream OCaml.

Yeah, its more just extensions to support their use cases at scale. Think of it more as bleeding edge ocmal, once they work out kinks/concerns they'll get merged back into the language OR if it remains ultra specific it'll stay in oxcaml.

Not a complete own version lol


Yes but the context of the thread is OCaml being "almost there". Having to build this stuff in-house is pretty good evidence.


I don’t know about that.

Python gets forked in other investment banks as well. I wouldn’t say that is evidence of any deficiencies, rather they just want to deal with their own idiosyncrasies.

See https://calpaterson.com/bank-python.html


Evidence of what?

The main user has been writing extensions to the compiler that they test before pushing for integration like they have done for the past twenty years or so. They publish these versions since last year.

Hardly a failure and certainly not something mandatory to keep things from failing over. Your initial comment is extremely disingenuous.


A different perspective is that JS has made practical application of PLT part of their secret sauce, and deepening into their PLT roots is thickening the sauce.


I thought that too (but more New York Christmas Eve movie impressions) until I saw steam coming from manholes in Denver. Blew my mind that it was a real thing haha.

In addition to the heating/cooling uses, the mint in Denver uses it to clean coins!


nope you have the article right. "if you copy an existing model, you can get a pretty comparable model!"

Really, [what the underlying paper](https://arxiv.org/abs/2501.19393) shows is how much you can improve existing models with several of the techniques released by openai recently (extending thought processes via "wait")

It shows we can get a lot of mileage out of that technique.


Surprisingly I find it does! The bulk of the time shouldn't be spent typing I agree, but often once I've mapped out what I want to do I need to type a lot. Improving my typing helped my "flow" as I made less mistakes and was generally faster once I got to the "make it happen" stage.


I was homeschooled from 2nd - 8th grade. My elementary school was trying to put my brother on adderall and my class had sorted me into the "blue" group of readers (colors of the rainbow for reading ability). I apparently came home talking about how I was slow and it was okay because we all learn at our own pace.

Definitely not a great school! both my brother and I ended up going to college and getting engineering degrees, and had zero issues with academics in high school. My mom did a pretty okay job but it was absolute hell on her, I entered high school ahead on mathematics/history but pretty behind on writing and science. The science I dont blame my mom for, all the curriculum at the time was insanely religious, so the ones we could find were very dry.


Those are sources, and while curiosity is great most kids are focused on specific things not everything. Kids need direction and somebody them to focus on things they dont want to learn- like a kid who loves animals isnt going to learn math or how to write well, and a kid whos interested in history might not care at all about science.

Parents are no better at this unless they are incredible focused on utilizing a curriculum and addressing their own issues along the way- And even then, learning with other kids is incredibly helpful. Talking to a computer is not a replacement for a teacher (yet).


> Talking to a computer is not a replacement for a teacher (yet).

I agree. The limitless patience and non-judgement of a computer is very valuable in a learning context. LLMs won't be better than the best private tutors, but its very likely they'll be better than 80% of junior high through college teachers.


Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: