tomcat27

This sounds like useful advice. Thank you.

I just can't think well on interesting problems while someone's talking to someone behind, someone at my face at the cubicle trying to hit a conversation. Hence, WFH.

Don't like team because finding good team 10xs everything but bad team 1/10xs everything.

It's way more likely the team is a misfit in real world unless of course there are no hiring constraints due to elitism.

Don't like to be a contractor.

> What companies want is team players who are often fun, light and friendly, actively work to build trust with a team through socialization, promote good practices and patterns and don't begrudge the fact that a portion of their day is to be spent in meetings, reviewing others' code and generally communicating.

It seems ok for a company to seek these from their engineers. I think, I want these too. I just don't like forced standups and sync up meetings. If someone did something cool and want to pull me aside to show? I'm all in. If someone wants help on something, I'll be there..

but it's just that the way things are managed is just dull.

let's do a standup meeting for 15mins where people explain - what they did yesterday - what they will do today - what are impediments

I propose an idea and the TL arbitrarily pulls off a blog post reference to shut down by saying "there's no business value".

umm. effing nonsense.

Firing over this very un-googly

I think it's fair to expect interviewee to solve 2 LC mediums in 45 mins given by the candidate. That's the current FAANG expectation. If a bad candidate is too costly for a company, so are a bad company too costly for a candidate.

People on internet seek engagement else why eve write a blog. Twitter is indisputably better for engagement.

They are not banks. They are forex traders.

It's not in the spirit of decentralization to stop withdrawals

I just don't understand why Apple Music search is so bad. Sometimes only exact match returns albums. It's so bad that appending "Apple Music" in google works better.

The problem is magic doesn't happen when people put jury duty on their CV

Not sure how that slipped. Sorry.

> You might mean 4 "doublings"

Rate of change: Enrollment increased by 4 times. Population increased by 100 times.

> How do you double elite universities?

Doubling elite universities: I mean more outreach from let's say T10 with next 10-20. This doesn't happen.

Ex: Top 5 won't cite papers written by someone in the middle. It happens but that's quite rare.

Ex: Faculty recruitment does not propagates from bottom ranking institution to top ranking institution. It's super rare.

Ex: Research collaborations are very clubby.

> It's more like we have an eliteness gradient (T10, T20, etc...) and the masses together decide what's "elite enough".

Masses opinion doesn't count. Eliteness is determined by elite. Consider an elite club exists at the beginning. Every 5 years new members are inducted into the elite club. Their opinion matters way more than everyone's simply because they are elite.

In most departments, the state of the art is a qualitative measurement. Something is great because a bunch of great people think it's great. There isn't a better approach anyway. But that comes with bad biases.

Ex: It's rare to have a Supreme Court justice coming from outside Harvard and Yale.

Ex: Some departments in Harvard would never hire someone as faculty just because they aren't from Harvard, Yale or a couple of other places.

The US population increased 100 times since 1800s from 5M to 500M. Harvard enrollment increased 4 times since 1800s from 1.5K to 20K.

The number of universities in the US did grew greatly during these 200 years, but the funding distribution to these universities is as skewed as the income inequality in the US. For instance, 60% of NIH funds go to <10 institutes in the country.

The scholarly work at these elite universities might not appear deserving of the preferential treatment from the funding agencies, if we start poking around reproducibility crisis, data hacking etc.

It seems fair to assume doubling elite universities would not be catastrophic to elitism.

Low barrier of entry is actually a good way to increase readers. I bet they are now not as influential as they were in 80s and 90s

It makes sense to write three versions on a topic.

1. A PPT Deck. A 2 mins elevator pitch, A 10 mins potty reading

2. A Blog post. Ok you got my attention, why should I care?

3. A Careful Argument. I see your point, explain me everything

People do this in ML research, and it's way better than other fields tbh.

> However, I think it's extremely unprofessional to write a blog post saying that your supervisor is superficial or your research group is performative art.

lol. You don't know the weight of crushed aspirations and a two year opportunity cost.

> I don't say this because I think we should just take the venerable professor's word for it, but because accusations in general, if they're vague enough, are impossible to defend against.

again lol. publication game is well know.

> I think if there's enough substance to make your comment specific, you should. Maybe even write a paper explaining why what they're doing is nonsense. If you're right, this would genuinely be a contribution to the field. If you can't be specific enough to be scientific about it, just keep it to yourself.

again lol. this is a very academic answer. Actually, I'm not sure university administrations write a paper on policies and ask for peer comments before making changes. They just do it, in the most brutal, authoritarian/top-down/corporate way possible. But the top brass needs papers from the bottom. Show me how much academia has been optimized by writing papers about academia. Well, even if they did they are useless.

How to be a Sam Altman: Find a Paul Graham

is gemini on similar situation?

I'm glad an insider is thinking about this.

But why do I have a feeling that it's written with an idea to make faculty life easy, not necessarily for a great student learning outcome, although that's what they say.

Alan Kay (Turing Prize 2004) argues against current standard CS curriculum. https://www.youtube.com/watch?v=PFc379hu--8

Anyway few unrelated thoughts.

---

The undergrad CS courses mostly do not deviate in content.The variance is in quality.

"pick good professor over interesting course" paid off for me in university.

Factors such as these matter: course pace, instructor and TA accessibility, course slides and other material quality, lecture interesting or boring etc.

---

Isn't it surprising that one of the most watched courses is by an YouTube tutor with a credential no american university will ever take seriously for hiring even an adjunct teaching faculty?

See Intro to Algorithms by Abdul Bari on YouTube

I'm cautiously optimistic about these style of interpretability works. They seem to tell us something without actually telling us. The ones in the past did not inspire to build better models. Translating a neural net into a decision tree makes interpretation no more useful tbh. Who knows! Do what you like. Good luck for everyone in finding these useful.

The goal of people doing basic research in math and science is not producing more economic value. They really do it just for fun. If what they do happen to be useful for others outside their club, great, but that's never the goal. They might twist words a little to get grants. Historically, their work has been useful. ;)

The problem at heart is research pace. Slowing the pace should help. i.e., it should be ok for faculty to produce half the papers they now produce and still retain their hopes on tenure, or whatever grants and glory.

I cannot get creative under pressure. I don't know how anyone can.

There is insurmountable amount of papers produced every year in AI, and everyone's in a rat race for some reason I don't understand. There is no way someone can exist in a department for more than a year without pushing a paper.

At least in AI, the courses students are taught, professors have an understanding are out of touch with the kind of mathematical rigor needed.

I can't speak for others, but I find scientific literature is not written general audience to understand, for a variety of reasons such as (a) publisher limitation on number of pages (b) to communicate core message right without distraction (c) to not bore a certain people etc.

Not to mention, depending on the sub-fields, at least in AI, there is quite a bit of noise in papers as they are explaining _their_ viewpoint which might differ from broad consensus or disproven in future.

I find papers are typically written keeping in mind a specific set of people, who are typically in that particular sub-field for some time. In sub-fields like quantum computation, that specific set of people authors have in mind while writing paper are <100, and those are not newbies.

Newbies can quickly get discouraged reading a paper out of the blue. Picking the right papers at the beginning requires exposure to the field.

It's not simple.

I really hope what you said is practical. I tried. Hope it's my fault because the alternative option doesn't exist for me.

sounds like a good plan to get tenure. lol

I'm probably in the same boat as those people, but in AI. Some of the problems I find interesting are just out of my depth. I don't know exactly what's required and from where to catch up to attempt a particular question in mind.

A PhD student's story on this. https://www.reddit.com/r/MachineLearning/comments/73n9pm/d_c...

How’s that solving leetcode att 30mins pace in 5 of 7 45mins rounds can get a candidate know about company. No wonder Blind is more vivid than interviews.

https://poop.is/composed Google.com

server crashed. lol.