THANK YOU! I studied AI in school, and it always bothers me when people think that LLMs are the only facet of AI. Between 2022-2024, I had a knee jerk reaction of explaining that AI is more than LLMs and that LLMs are really a small subset of the entire universe of AI, yadda yadda yadda. Now I’ve given up and roll my eyes as someone tries to tell me about the cool new Claude skill they built.

What’s funnier is people think I hate LLMs. That couldn’t be further from the truth; they are a fantastically interesting and innovative technology! “Attention is All You Need” is a great paper, and super impactful. I just hate that people are outsourcing their thinking to a chatbot and neglect the rest of my field of study.

The job market in tech is not good right now, so they may not. I have a friend who is a dev and was getting heavily recruited by Palantir because he had a clearance already. The money was really good, and he was at the cusp of taking the job because he had been ghosted by other companies for months and only had a handful of interviews after being laid off six-ish months ago. Luckily, another place gave him an offer for a bit lower compensation, which he took. If that other place didn’t come through, though, I think he would have gone with Palantir so that he could get some income.

ETA: That said, Palantir is evil, and my friend was in that position because of the BS that is the US economic system, which is perpetuated by the people at Palantir.

The “well rounded person” shit is only ever given as a justification for forcing STEM majors to pay for liberal arts courses. I’ve never seen it go the other way, and it should. For every credit hour a STEM major spends in a humanities course, a liberal arts major should have to spend in a technical course.

Absolutely! I say this as a Comp Sci major who loves the humanities (and almost studied History). General education should encompass introductions to both the STEM and humanities areas. It is equally frustrating when I can’t walk through a 7th grade level algebraic function with someone with a Master’s degree in International Relations and when I can’t reference a fairly common part of mythology with a software engineer.

I wouldn’t go as far as to say it’s only a small percentage that are real degrees these day but it’s definitely lower than it should be.

I agree. I think a lot of degrees are still real degrees, but the entire ecosystem has been degraded to the point that quality across the board has diminished. So, the most “rigorous” degrees now are equivalent to a run-of-the-mill degree a generation ago and so forth. Ultimately, the run-of-the-mill degrees of yesteryear are now just diploma mill degrees.

I hate to say it, but a lot of it is e-learning and online degrees. It’s a lot harder to engage with material, with a class, or with the professor themselves behind a screen hundreds of miles away. Even when you put everything into the work, it still just is not as engaging because you don’t have the same dynamic because you can’t just drop by your professor’s office for office hours or get the same level of help or group learning. In undergrad, I used to help others in my classes, and vice-versa, while also going to office hours to clear up details. Online, if it’s not impossible, it’s at least orders of magnitude more difficult. So, the quality of learning drops a ton.

If I go back for another Master’s or a Doctorate, I will only do in person classes.

There are a surprising amount of false positives when using object detection on maritime imagery. While a carrier is a spec, there are a ton of specs in the ocean that can look similar enough. Plus, weather has a huge hand to play. If it were always perfectly clear, then it’s an easier problem, but one cloud can really mess up the detection. Ultimately, ship detection is a difficult problem (not intractable but still hard).

Sort of. Satellite resources are surprisingly scarce, so a lot are focused where people are, i.e. land. Plus, for the imagery sats that are focused on the ocean, ships are also tiny in a literal ocean of blue. It’s just a spec. While the resolution could be good, have fun looking for that spec. That’s why most countries use signal collection to locate vessels at sea. (I’m over-simplifying a lot, but you get the picture)

Even moreso with ad-blocker and PiHole. I swear that every corporate website has me do a captcha of some kind every time I click a link.

On that note, just denying trackers breaks SO many email links, and it feels like it gets worse and worse each year.

However, many times getting a carrier unlocked phone carries a price premium.

That’s a good point. I’ve started advocating for buying phones lightly used (1-2 generations behind). Until just a couple of months ago, I was rocking a phone from 2019 with no issues. When I upgraded recently, I bought a Pixel 9 from a reseller selling one with an open box and a slight scuff on the bezel (that gets covered by a case anyway). Now I have an almost new phone that works like a dream for almost 75% of what I would get buying it directly from Google.

Sounds perfect for some trust-busting action. If only we had a capable government…

If God’s mind were a soup of linear algebra doing stupid number tricks, then sure (with the assumption we’re just talking about LLMs).

In reality, no.

Source: I study AI and work on it professionally.

100% agreed. It should not be used as a replacement but rather as an augmentation to get the real benefits.

I study AI, and have developed plenty of software. LLMs are great for using unfamiliar libraries (with the docs open to validate), getting outlines of projects, and bouncing ideas for strategies. They aren’t detail oriented enough to write full applications or complicated scripts. In general, I like to think of an LLM as a junior developer to my senior developer. I will give it small, atomized tasks, and I’ll give its output a once over to check it with an eye to the details of implementation. It’s nice to get the boilerplate out of the way quickly.

Don’t get me wrong, LLMs are a huge advancement and unbelievably awesome for what they are. I think that they are one of the most important AI breakthroughs in the past five to ten years. But the AI hype train is misusing them, not understanding their capabilities and limitations, and casting their own wishes and desires onto a pile of linear algebra. Too often a tool (which is one of many) is being conflated with the one and only solution–a silver bullet–and it’s not.

This leads to my biggest fear for the AI field of Computer Science: reality won’t live up to the hype. When this inevitably happens, companies, CEOs, and normal people will sour on the entire field (which is already happening to some extent among workers). Even good uses of LLMs and other AI/ML use cases will be stopped and real academic research drying up.

this “cameras for everything!” idiocy.

That’s why I’m so impressed with how well it’s actually working. When they get off that really weird self-imposed restriction, it could be an interesting technology.

Not great performance at all.

That’s better than I was expecting to be perfectly honest.

I’m pretty impressed with the technology, but clearly it’s not ready for field use.

It’s also not all-or-none. Someone who otherwise is really interested in learning the material may just skate through using AI in a class that is uninteresting to them but required. Or someone might have life come up with a particularly strict instructor who doesn’t accept late work, and using AI is just a means to not fall behind.

The ones who are running everything through an LLM are stupid and ultimately shooting themselves in the foot. The others may just be taking a shortcut through some busy work or ensuring a life event doesn’t tank their grade.

I see both points. You’re totally right that for a company, it’s just the result that matters. However, to Bradley’s, since he’s specifically talking about art direction, the journey is important in so much as getting a passable result. I’ve only dabbled with 2D and 3D art, but converting to 3D requires an understanding of the geometries of things and how they look from different angles. Some things look cool from one angle and really bad from another. Doing the real work allows you to figure that out and abandon a design before too much work is put in or modify it so it works better.

When it comes to software, though, I’m kinda on the fence. I like to use AI for small bits of code and knocking out boilerplate so that I can focus on making the “real” part of the code good. I hope the real, creative, and hard parts of a project aren’t being LLM’d away, but I wouldn’t be surprised if that’s a mandate from some MBA.