mirawelner

I worked for professors for free in undergrad, and then later for money also in undergrad.

Profs are always looking for research assistants. Granted now is the wrong time to try to break into the industry given the NIH cuts. I was able to get a contracting position with university of pittsburgh and a free gig with CMU just by asking and that pays the bills but its not too much money. However if you don't already have publications you may have to wait until this chaos has wound down (honestly even if you do have publications it'll be hard to get paid until the chaos winds down I think I mostly just got very lucky).

You can always get publications by working in a lab for free and have a real job simultaneously but that is a lot of work.

There are also startups that have similar job descriptions. I worked for Signature Diagnostics for a year and a half and it was very bio/science heavy. But the key skill to break in is publications. I personally don't think this is a great metric, but unfortunately it is the metric.

Personally I've had a lot of success working in non software science/engineering labs. I'm currently applying software and AI to biotech (don't worry its not LLMs). I think that is where a lot of the utility of AI is because we 'traditional software' is quickly becoming a solved problem. Its a lot more interesting in my opinion to use software to solve real science and engineering problems that they haven't been applied to, and it typically pays well unless you work directly for a university (in which case you may get defunded because of the current administration and even before the current political situation the pay was lousy)

this reminds me of the (now discontinued) lambda tensorbook that I bought a few years back. AI computers are interesting because they are extremely similar to really good gaming computers but the marketing (and some of the hardware) is different.

I don't fully understood why they are separate from gaming computers other than marketing, but I am very happy with my tensorbook. I got it when I was very young and impulsive I don't know if I would get it again, although that fact alone speaks to the computer's longevity.

This was also the average college student yesterday and it will be the average college student tomorrow.

I am working on creating a GAN for predicting the next few heartbeats given an ECG signal and hoping that the administration does not defund this project because I am being paid by a university project funded by the NIH.

It brings me so much happiness that there is enough overlap between software people and jazz playing people that this is on the front page

I have also found LLMs (in my case I use deepseek) to be really good at HTML/CSS, also probably because I'm not great at it.

But yeah if I'm going to search stack overflow, that is when I use LLMs because my output is going to be bad anyways, so it might as well be bad and fast.

I really like the comparison between talking about politics at work and talking about salaries at work. They are both things that were just 'not allowed' for vague reasons and they both turned out to be really important things to bring up at work.

I have a speculation which may be hubris that cursor is such a hit because its better than a lot of programmers because a lot of programmers aren't very good.

And then when an actually talented programmer tries to use cursor it is much less good then they are alone.

But idk maybe I'm being an arrogant snob who underestimates the vast majority of programmers.

I’m working at a biomedical lab at UPitt full time and another biomedical at CMU in my “free time”(lol) and it’s bizarre how it’s now totally okay to talk about politics at work. People are sharing political articles on the teams chat.

Every week I hear about a new solution to plastic. Every week more microplastics end up in my brain. I feel like we are not solving the problem.

Im definitely in the first camp which explains why im going the phd route and why I am poor :)

The cool thing about people showing off their webdev skills is that the portfolio site itself is the demo. I know this isn't a portfolio site but it reminds me of my webdev friend's sites. I love when people push web navigation to its limit

AI that is smarter than the average human has existed for a while. The average human isn't that bright. If we have an AI smarter than a smart human that would be impressive. But honestly average human hallucinations are worse that LLM hallucinations at this point.

(Sorry I work in a ML+Bio lab in the US and I'm depressed and angry)

I spent a year working at Purdue, and while I never encountered the co-op program that sounds very on-brand for Purdue and a very cool project!

Good for your sister!

THANK

GOD

Edit: switched

I have always wondered if the threshold for 'number of papers published' to get a professorship is higher than the number of high quality papers that can reasonably published in that amount of time. So, people cut corners and publish bad papers or commit outright fraud. Because they do that, this keeps the required number of publications unreasonably high.

I’ve been meaning to switch to DuckDuckGo for years - this would be the final straw for me

Beware for if you continue down this road you will end up sitting in class taking notes in markdown… yes I did do this… I am afraid I am beyond salvation

I like lobste.rs. It gets heat from being invite only and therefore somewhat snooty but I’m a big fan

If you are going to be funding boss spyware at least have the guts to face the backlash

There is also something to be said for the fact that good researchers tend to enjoy weird research and we can attract talent by encouraging weird research

People resorting to protest via causing inconvenience. Which is probably one of the more effective forms of protest anyways.

The main reason I blog is because I work (or am starting out working in) academia and therefore I have to write papers. To write academic papers you have to write terribly, using passive voice whenever possible.

I blog so that I know I am still capable of writing coherently, rather than in horrible academic language.

I don’t think it is - the whole blog is in txt

https://misc.l3m.in/txt/

Sometimes I think to myself 'I should ignore HackerNews and do all my posting on lobste.rs' and then I come across a post like this and I remember why I love this site.

While I disagree with Ed Zitron's opinion that we are at 'peak AI', I do completely agree with his other points, especially the points he was making that OpenAI and Anthropic had no incentives to make things more efficient and therefore they were just drinking from the money hose. If anything it is shocking that nobody had done what DeepSeek did earlier. Good for them.

I think at the end of the day the reason that the opensourced DeepSeek is because they are programmers. Programmers like to show people the cool stuff they did. I had a boss who was rich enough to retire but was working three jobs because programming is cool and fun and he wants to do cool and fun things and show people the stuff that he did.

Everybody is trying to come up with a money related reason for why they open sourced it but at the end of the day the people who made it are engineers and not buisnesspeople. DeepSeek is really freaking cool, and they wanted to show people the cool thing they did.

I found that people think that whatever you are doing must be extremely cool if you display it with the 3rd dimension being time. So create a video of each frame of the hyperspectral image and it shifts from one frame to another. Idk if it is super useful but if you put the video on a powerpoint it gets lots of excitement