this post was submitted on 20 Jun 2024
40 points (100.0% liked)

TechTakes

1427 readers
246 users here now

Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] V0ldek@awful.systems 16 points 5 months ago* (last edited 5 months ago) (3 children)

The reason Lambda School's hiring rates dropped should be obvious to anyone who spends time around bootcamps, or education in general. Lambda School tried to scale up. Staff warned about a downward trajectory back in 2018, in an internal memo that said, “Placement to date has been manual and one-off, which isn’t possible to scale.” That was back when they were training a few hundred students a year, and at their peak, they trained 2,700 students.

This is hubris beyond my comprehension.

I completed my BSc. (3y programme) and MSc. (2y programme) in Computer Science at University of Warsaw. It's a really good programme, by far the best CS programme in all of Poland, university in the capital (largest city as well). Publicly funded, but very successful in research, so our staff has many ERC grants which pay out a lot, so it's probably one of the better funded ones as well. Population of Poland is about that of California.

Yearly, less than 200 students are enrolled for the BSc at UW. On MSc. this is about 100. So you give lectures to around 150 students in the biggest courses. A single class or lab is led by a TA for 15-20 students. In person.

I'm not saying any of this to be elitist or some shit, but just comprehend the scale we're talking about here. This is the cohort size that is manageable by institutions that know how to do this, have experience, staff, and funding. The bottleneck in this system is staff - you simply cannot have more students without hiring more than the couple dozen professors already there.

When you say "they scaled up" I though we're talking they were enrolling dozens and scaled to a couple hundred. 2700 students per year is the scale of the absolute largest universities in Europe, backed by both public funds and institutional investors. How the fuck do you expect to give anyone any education at that scale with online classes and no TAs? That's insanity. Like, "I'm going to build a spaceship in my garage with a box of nails and $100" level of insanity. What do those people think education even looks like?

[–] gnomicutterance@awful.systems 14 points 5 months ago

One of the things that happened during the Great Low Interest Rates Decades is that it seems like anyone who fit a certain profile (millennial white guy with american citizenship, a computer, and at least a modicum of what passes for charm among the nerd elite) could convince both VCs and the US government that there was tons of money in disrupting the delivery of some legacy sector of society. Sometimes they were correct (eg. buying stuff without going to a retail establishment), sometimes it seems like they should have been correct and yet somehow have failed to make money anyway (Uber), mostly they were comical (Juicero). But the ones that are the most excruciating are all the places where you really, really can't frictionlessly deliver at scale, because large-scale human intervention is necessary: education, health care, customer service.

The promise of the American tech boom is massive online delivery without people. Employers hate their employees, and government is always willing to be told that doing without employees is industrial progress.

[–] sailor_sega_saturn@awful.systems 12 points 5 months ago* (last edited 5 months ago)

What do those people think education even looks like?

The same think all coding bootcamps think education looks like: try to ~~take over the world!~~ poorly teach fullstack web development with none of the fundamentals!

[–] skillissuer@discuss.tchncs.de 3 points 5 months ago (1 children)

So it seems like you're my former neighbour, then. Depending on how you count, UW might look like it's both overstaffed and underfunded. This is because i'm mildly certain that most of these TAs aren't hired as TAs, because nobody really pays them. All PhD students, definitely all at Kampus Ochota, have a requirement to take a certain number of teaching internship hours, which is about 40-60 per year depending on field. (this is, btw, anomaly on national scale and from next year or so new admissions will be paid for these TA hours).

At my faculty (Chemistry; about 220 admitted for BSc, about 80 for MSc annually) that's enough to man almost all the more entry-level labs (that require more supervision), along with a proper salaried PhD or sometimes two. (We mostly make sure that students don't do any stupid/dangerous shit, most labs are 12-15 students, first year labs are 30-40 and are limited by size of labs. We're talking about 1 staff, incl TAs, per 4 to 6 students) I'm not sure how MIM makes use of that human resource, but wouldn't surprise me if some classes are just led by some PhD students. This is definitely not a resource that Lambda has, so actual costs if anything should be even higher

ERC grants, haven't seen many of them. European Regional Development Fund, a few grants sourced from Norway and Switzerland, NCBiR, NCN, FNP, some funding straight from ministry, there's more of this. MIM is a notch or two higher than most of faculties so i don't doubt it, but these grants are supposed to go for research, not for teaching. There are other sources of funding for that, and especially after new minister of science was installed, that money became available (it's a special hell to navigate all of this)

That said, our faculty hasn't been saved from bullshit hype. There's nanostructure engineering course now, launched some three years ago or so, which is not engineering degree and mostly deals with bulk nanoparticles and such, and this is something that simply does not have much use (some sensors, limited use in catalysis but it has problems with reproducibility and dies quickly). Nano- hype happened mostly in early 00s and started dying around 2015 or so, so it's not even right on timing. This deters few however and you can find loads of bs papers with shit like catalyst based on chitin nanoparticles made from recycled sustainably sourced beetles or whatever with non-reproductible results published in MDPI or Frontiers. There is, or was, "blockchain and AI lab", tucked somewhere at CeNT (i take that yall's have thrown them away from MIM and they inhabited first hospitable room since)

[–] V0ldek@awful.systems 6 points 5 months ago

his is because i’m mildly certain that most of these TAs aren’t hired as TAs, because nobody really pays them. All PhD students, definitely all at Kampus Ochota, have a requirement to take a certain number of teaching internship hours, which is about 40-60 per year depending on field.

Depends on the course, but most TAs are PhD students, however some of the labs are also done by M.Sc. students hired on a contract (I was one!). PhD students don't get paid explicitly for teaching hours, it's part of their duties, but

but these grants are supposed to go for research, not for teaching

one of the main uses of grants is to hire PhD students! So having funds from research directly impacts the quantity and, perhaps more importantly, quality of PhDs, since ERC grants are very prestigious.

The tragic state of PhD students being on starvation salaries is a fact, but it's a wide systemic problem with the entire country's education. That being said, even with this "cheap labour source" you still can't run courses for thousands of people!