What's interesting is that if people put "generative AI" many people seem to read it as actually making sense.

Basically, the idea is the same as if you were going to the gym and while there paying people to exercise for you. That would sound stupid to people. But because doing the same thing with learning and AI has buzzwords in it, it sounds reasonable to many.

...but that's a pretty low bar.

Or at least take a math or stats class. Comparing "a third of" thus something like median for a starting salary, to the "average" and hence mean-like thing for the salary throughout the life. If you want to compare apples to apples, compare the same sort of averages for the same age groups.

See e.g. https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.forbes.com%2Fadvisor...

Among the same age group (25-34), median salary:
High school only: $36.6k
Bachelors: $59.6k

$15k difference per year.

If you want the median salary for college graduate for 25-retirement age group you get $75.2k. The median high school only salary in that range is $45.6k. So that's the thing you want to compare to that 44k mentioned in the idiotic article. Basically as you get older, sometime mid-career, the difference grows to about 30k.

Over the course of a lifetime, a median college graduate earns 2.8 million as opposed to 1.6 for a high school graduate. So over a lifetime the median person gets more than 10x the return for the investment.

Is everybody going to make this sort of salary? No. These are medians. But college is an extremely good investment for more than half of people. The number of people for whom it is a poor investment is actually rather small (though they do of course exist). Even the worst majors have median starting salaries above similar age group median high school only salaries.

Sorry, misread that, not 137 (it did seem low). It is more than 2500.

Not just this, he was also a coauthor on 137 scientific articles, which were cited 239852 times (per scholar.google.com). On the other hand, we have the opinion of some "anonymous coward" on slashdot.

In this case, they rediscovered X windows. There is no new revolutionary idea behind this. This has been done in principle many decades ago. I used to load netscape through X running at the department server at school. I didn't really do it for the sake of network/CPU speed though, I did it so that I could easily use the library. So what they did is a new implementation of a very old idea. While it's not a bad idea in principle, I doubt it will catch on. Fast network connectivity is still not ubiquitous. A local browser will work much better in case of slow connection on relatively simple pages.

Yes, CS students are the smartest there are. There is nobody smarter and more passionate on the planet. Anybody that says otherwise is jealous.

You!

I teach mathematics to (among others) CS majors. I do not share your optimism.

Since when are MIT and Stanford in the Ivy league? Actually Cornell is the only mentioned Ivy league school.

And ... my my are we offended today. Somebody said something that could be interpreted as programmers not being the smartest people on the planet. Oh the humanity.

By the way, have you been to China? Have you tried to breathe in Beijing? Actually have you even tried looking across a street in Beijing? If they took our pollution, they can have it. Also, it didn't seem like they were so much wealthier on the whole.

Ahhh exactly the example of a perfectionist. The same person who's probably willing to pay via credit card at a supermarket. Typical paranoia. Either everything or nothing. When it's snowing outside I'll put on whatever clothes I have, even though they are not designed exactly for the weather. Yes I might be a little cold after a while, but oh well. You on the other hand will either not go outside at all or run around naked, because if you dont have everything designed for exactly the right conditions, you might as well not put on any clothing at all.

Look, there are two thing, encryption and authentication. Don't conflate the two. Encrypted connection is for protecting against different things than authenticated connection. Saying that you can't have one without the other is stupid. There's no reason to ever send anything cleartext. Yes, it might be better to authenticate, but it is not all that difficult to obtain a certificate for a domain if you can control the domain for a bit, which is exactly what you need to impersonate a site that google would be connecting to. The thing is as long as you have any certificate given for that domain, then you're Bob.

Authentication and Encryption are TWO DIFFERENT THINGS.

Also this is google we are talking about. They FOR YEARS could not cobble a two factor or one time password authentication together. So I won't take any lecturing about how concerned they are about security.

BTW, google has been using quite a bit of software I wrote, for free, even android apparently used a bit of my software as my website pops up in their license files. I don't feel at all bad for them giving something to me for free. Plus it's not free. They are not doing it out of the goodness of their hearts (you did notice those ads in gmail did you not?). They are making quite a bit of money. Enough to not pay taxes on lot of it and make lots of people angry.

If you only do these things once a year at most you forget how everything works. The documentation is terrible. So if you understand how to set everything up and what all the acronyms are. I am a mathematician and I understand the mathematics of public key crypto really well. I don't know the little pointless details. Yes onc eyou know what to do and what to get and where to install it etc... yes it's not that much busy work. But rereading all the documentation every year (yeah understanding what e.g. "CSR" is, is pointless unless you do it often).

It's not just startssl, it is the combination of badly written documentation starting with the pop3 server, through the doc on startssl. The problem with these docs is that they only explain anything once you know exactly what to do. At which point they are useless as well.

Plus the startssl website kept freezing. It's an incredibly badly designed UI. very intolerant of pressing the wrong thing for example.

You know, I'm not as smart as your regular user I guess. I only have a phd. Dropped out of school after that.

Yeah, apparently no security is better then some security according to google.

Got certificate from startssl, but it's a pain. Couple of hours of totally pointless work.

Another example where perfection is the enemy of good. This is my gripe with most computer security people. One of the reasons why encryption is not as widely deployed as it should be is this attitutude that "it must only be perfect".

I think a combination of tools might be the answer. I use maxima (wxmaxima frontent) when I need a cas. I use my own software Genius when I need to compute something numerically, and I often use it for in-class demonstrations (I often end up implementing whatever it is I need at some particular point). I can't remember when I last used octave, but that also sometimes happens when tehre's something genius can't do. I tried to make the interface to genius friendly, though of course there's always plenty of room for improvement. Generally it's a "command line" type interface, but I think it can do some pretty graphs. Too friendly tools generally end up being not very flexible. So it is worth it to spend a bit of time learning the less friendly ones.

By the way, I am getting ready to make a new genius release this weekend, I have just one more thing to do on my list before a release.