I know LLMs are getting the majority of the attention of late, but there are others of us training and using AI to…
…reads MRIs and video to detect cancer
…analyze genomics for early target discovery
…assisting surgeons
…folding proteins
And the list goes on in other fields as well. Just hoping the recent AI counterculture doesn’t stigmatize other uses of AI.
Am I a fan of Claude code? Not particularly, but I have used it on occasion. And I’ll never understand someone using an LLM to write anything (especially a comment on a site like HN) intended for consumption by other people. Not because I think it’s subpar, but because the point - IMO - is to make human connections, learn, teach, and debate. That’s hard (impossible?) to do if you’re just typing a 30 second prompt and then copy/pasting the output.
Just wanted to note that this is how I work. I rarely have any window full screen/maximized and hate it when a website or application is built assuming a giant monitor with a maximized window.
I’ve never found a setup with multiple desktops or similar with a way to quickly switch between apps I’m using more than “editor slightly more left, browser slightly more right, …” and just clicking on a border I know brings that app to the front. I’m sure many think I’m crazy. That’s ok. :)
That said, I generally hate the new OSX UI. Every UI element that is non usable just became larger and wastes space I should be able to utilize. Likewise, it made some operations insanely frustrating (here’s looking at you, corner drag resize!).
In my 25+ years programming, I don't yet think I've met a programmer who hasn't dreamed of "just being a lawnmower" (or some equally equiv, "mundane" job). Just the idea of ending the day having accomplished something and being able to 100% switch off work and just relax. The same is probably true of other professions like lawyers and doctors.
AI is definitely going to change things. It already has, but I don't think what it's done is the real change coming. Right now it's just a new tool in the toolbox. It does great at some things and terrible at others. Note: I'm speaking purely about LLMs/generative AI, because other forms of AI are -incredibly- useful and have been for a while (think DeepFold). I only mention it because it's important people remember there's different types of AI models used for all sorts of different things.
What's important to keep in mind are a few things:
* We all stand on the shoulders (abstractions) of giants. People of the past lamented about C, Java, ... every abstraction that came before. Did those abstractions cause the next batch of programmers to lose knowledge about what was really happening? Sure, but each also enabled an explosion of new ways of thinking and problem solving that brought us to where we are now. If you can recognize that, then that will help.
* You became a programmer/engineer for reasons (what I'll call "itches"). Only you know what those are. Buy, my guess is that - like most of us - there was something about the -problem solving- and logical analyses/thinking taking place, and seeing others gaining benefit from that work. First, if you look to another job, pick something that scratches those same itches, otherwise you'll be bored and depressed again very quickly. More importantly, the reason you're even asking the question is because something about AI feels like it's taken that "scratch" away from you. Keep in mind that your "itch" can be abstracted, too. Instead of solving a (for example) network byzantine problem, go one level higher, which has the same problem at the human-scale. Now AI becomes a tool to help you come up with and solve the problem instead of one that's replacing you. Same itch, different level, different tools.
* The comment mentioning "shadowing someone" is some solid advice. Right now you're succumbing to a "grass is greener" effect. That will disappear quickly if you do switch, and you'll still be in the same place if those itches are not being scratched. Again, identify the itch first.
Finally... and I can't stress this one enough... whenever I've been down due to the "itch" not being scratched at work (for any reason, not just AI), I've personally found the best solution to the problem is to volunteer and help others. This could be anything: food bank, big brother, translation services, helping to teach your native language to someone else, or even teaching something you consider important (i.e., programming) to kids at your local school via an after-school program. Every single time I do this I walk away feeling great. Inevitably there was a single point in the day in which I dramatically affected one person's life positively, and they - in turn - affected me similarly. In a class of 10 kids who want to learn to make games, there was 1 who had no way of being able to do this on their own due to at-home problems. At a food bank it was an elderly woman who couldn't speak English and I happened to be the one person who could speak her language (she was so excited to talk to someone). Or being in the children's ward at the hospital over x-mas, delivering toys. And sometimes I'd even walk away with a new set of ideas where my programming skills could solve a new set of problems I discovered existed and lead me to a whole new career.
I’m no expert either, so I hope one can corroborate or correct me…
My understanding though is that these steps are really the very beginning. Using a quantum computer with quantum algorithms to prove that it’s possible.
Once proven (which maybe article this is claiming?) the next step is actually creating a computer with enough qubits and entanglable pairs and low enough error rates that it can be used to solve larger problems at scale.
Because my current understanding with claims like these is that they are likely true, but in the tiny.
It’d be like saying “I have a new algorithm for factoring primes that is 10000x faster than the current best, but can only factor numbers up to 103.”
Great post. I feel obligated to reply with a similar post a friend wrote a while ago that’s probably made the rounds here as well: https://prog21.dadgum.com/116.html
> Any app that ever claimed to tell you what "Hemingway would say about this blog post" would evidently be lying — it'd be giving you what that specific AI model generates in response to such a prompt.
First, 100% agreed.
That said, I found myself pondering Star Trek: TNG episodes with the holodeck, and recreations of individuals (e.g. Einstein, Freud). In those episodes - as a viewer - it really never occurred to me (at 15 years old) that this was just a computer's random guess as to how those personages from history would act and what they would say.
But then there was the episode where Geordi had to the computer recreate someone real from their personal logs to help solve a problem (https://www.imdb.com/title/tt0708682/). In a later episode you find out just how very wrong the computer/AI's representation of that person really was, because it was playing off Geordi, just like an LLM's "you're absolutely right!" etc. (https://www.imdb.com/title/tt0708720/).
This is a long-winded way of saying...
1. It's crazy to me how prescient those episodes were.
2. At the same time, the representation of the historical figures never bothered me in those contexts. And I wonder if it should bother me in this (LLM) context either? Maybe it's because I knew - and I believed the characters knew - it was 100% fake? Maybe some other reason?
> Seasoned Rust coders don’t spend time fighting the borrow checker...
Experienced Rust coders aren't going to find themselves in various borrow checker (and lifetime) pitfalls that newbies do, sure.
That said, the borrow checker and lifetime do cause problems for even experienced Rust programmers. Not because they don't understand memory management, lifetimes, etc. But because they don't - yet - fully understand the problem being solved.
All programs are an evolutionary process of developing a solution to a problem (or many problems). You think one thing, code it up, realize you missed something or didn't fully grok the issue, pivot, etc.
Rust does a great job in the compiler of letting the user know if they've borked something. But often times a refactor/fix in C/D/Zig due to learned (or new) requirements is just a tweak, while in Rust it becomes a major overhaul because now something needs to be `mut` or have a lifetime added to it. I - personally - consider that "fighting" the borrow checker, regardless of how helpful (or correct) it also is.
I’m sure your question was rhetorical and sarcastic (that the app exists makes me sick).
It absolutely is not. AI counts as a medical device: it’s used to help diagnose and inform medical treatments. However, this is a loophole, because (quoting fda.gov):
> FDA does have regulatory oversight over devices intended for animal use …. Pre-market Approval is Not Required: The FDA does not require submission of a 510(k), PMA, or any pre-market approval for devices intended for animal use.
The FDA will only step in after complaints or enough pets start dying.
I've personally found that the time taken to think through a discussion is akin to an inverse guassian curve:
- on the left tail are people who know little-to-nothing about (or have little experience with) the given topic and neeed a chunk of time
- then as knowledge and experience increases, less time is needed, eventually peaking out at what appears to be instance understanding + ability to communcate effectively about it
- but then something interesting happens when they get even more experience + knowledge: they now know about all the edge cases, things that go wrong, etc. and once again take more time to think through the topic
I've also found that most everyone is the same in this regard. Every once in a while (like any normal distribution) there's an outlier on one side of the spectrum or the other, but for the most part, everyone is the same.
Where people tend to differ is in their coping skills in such situations. Early in my career I had to learn to ask people to explain their thinking. Later it was me slowing down and realizing there's likely more to it than I think (and for those behind me).
Now it's me telling those at the peak of the curve to slow down, because while they may be right, and -maybe- they've thought it through, that's probably not the case.
TL;DR to anyone who thinks they are a slow thinker - you probably aren't (like imposter syndrome), and just need to learn to slow the room down. Doing so will help you, others behind you, and those in front of you.
First, this is great information in an area I know very little about.
But I’m curious - from your experience - how do you know the OP isn’t pretending in order to learn about new avenues to block or attack or to track down people who are trying to circumvent?
I don’t mean that as a “be careful”. You’re the expert compared to me and for all I know these are unblockable. Or maybe those doing the blocking would already know about them? So I’m interested in just understanding more.
Regarding the former, I’m nearing retirement age, so personally I don’t care as much; I’m no longer “investing [in] a dead craft”. Assuming it is dead (I don’t think it is).
Re the latter, I have rejected it. I love problem solving. And I consider programming a tool I use to solve problems. Regardless of whether it’s an LLM or my old C text book, if I limited myself to only what came before me, then I can’t possibly improve on the current situation. My solutions would be in a perpetual state of stagnation. I can’t speak for others, but that sounds boring AF to me.
By that logic, if modern LLMs existed in the 80s, you’d have never learned Haskell, Ocaml, Rust, Go, Erlang, … and all the cool concepts and ideas that came with them. You’d still be programming Basic and Fortran, simply because that’s all the models knew.
AI may be helpful at times, but to limit one’s self to only the knowledge and experience they have is… short sighted at best.
I haven’t lived in NYC, but I have lived in Boston. Isn’t the real concern winter? Has Waymo (or any other self driving tech company) shown that it can handle the snow well: non-visible lanes, downshifting to avoid braking, etc.?
This feels very similar to the “radio” or “restaurant” problem:
You’re driving down the street trying to decide which restaurant to stop at (or scanning through the radio trying to decide which song to stop on).
If you stop at the first, there’s a good chance something better is ahead. But if you wait too long then you risk getting stuck with something you don’t really like (the problem assumes you can’t go back).
If I remember correctly, mathematically you skip the first 1/3, but keep track of your “best”. Then stop at the next option that’s >= than your current best or maybe the next thing you like.
With respect to skis, I have the same issue every year with a ride on lawn mower. Do I just pay someone weekly or buy one outright and do it myself? In this case I loathe mowing, so I don’t mind paying. But with skis it’s a question of just how much I’ll ski after this stretch, regardless of whether or not this stretch is 1 or 20 days. Because there are additional costs (and benefits) to ownership beyond the initial purchase.
I think the distinction (from the OP’s perspective) is that it’s “just an app” to the end user.
When I download the Slack desktop app, I might know it’s running on electron because I’m a programmer. But I didn’t have to download and install nodejs or anything else. As far as I’m concerned it’s a completely self contained application.
The end user needn’t download and install Janet or even know that Janet exists.
That can obviously be done with any interpreted language, but Janet makes it really simple out of the box.
…reads MRIs and video to detect cancer
…analyze genomics for early target discovery
…assisting surgeons
…folding proteins
And the list goes on in other fields as well. Just hoping the recent AI counterculture doesn’t stigmatize other uses of AI.
Am I a fan of Claude code? Not particularly, but I have used it on occasion. And I’ll never understand someone using an LLM to write anything (especially a comment on a site like HN) intended for consumption by other people. Not because I think it’s subpar, but because the point - IMO - is to make human connections, learn, teach, and debate. That’s hard (impossible?) to do if you’re just typing a 30 second prompt and then copy/pasting the output.