8. Links #3: the real AI was inside us all along

134 words

Olivia Solon: The rise of ‘pseudo-AI’: how tech firms quietly use humans to do bots’ work

It’s hard to build a service powered by artificial intelligence. So hard, in fact, that some startups have worked out it’s cheaper and easier to get humans to behave like robots than it is to get machines to behave like humans.

Brian X. Chen and Cade Metz: Google’s Duplex Uses A.I. to Mimic Humans (Sometimes)

In other words, Duplex, which Google first showed off last year as a technological marvel using A.I., is still largely operated by humans. While A.I. services like Google’s are meant to help us, their part-machine, part-human approach could contribute to a mounting problem: the struggle to decipher the real from the fake, from bogus reviews and online disinformation to bots posing as people.

Is a PhD right for me?

575 words

When I was almost finished with my studies, I thought for a long time about whether I should do a PhD. On the one hand, I was in actual love with my master’s research topic and I had the opportunity to continue right where I left off. On the other hand, PhD students lead a stressful life of imposter syndrome and are paid a pittance.

I spent quite some time searching for articles and videos with titles like “why I quit my phd” to see if any of their feelings and thoughts sounded like the sort of things I might think. That was a useful activity which I would recommend to any prospective PhD student.

Right now, I am 2 years into my PhD.1I have heard multiple colleagues say that this is the worst part. Long enough in to think you should have visible results by now, but not long enough to actually have them. The title question remains. Here is a list of things that would have allowed me to make a more informed decision.

  • Imposter syndrome starts the day you’re hired and won’t stop for many years to come.
  • It is normal to fantasize about quitting every so often.
  • You will at some point realize how much more money you would make in industry. That is how much you’re effectively paying to do a PhD.
  • You will like your job less than you expected beforehand.
  • It is probably better to drown your sorrows in online procrastination than in substance abuse.
  • Free weekends will be rare, or at least you will feel guilty about them. Other people will always seem to be working more than you.2This is partly sampling bias. You do notice when your advisor sends you an email at 11pm on a Sunday, but you don’t notice when she doesn’t.

    There are good reasons for working outside office hours. Using your brain for 8 hours straight is hardly possible. Working less during the day and then a bit in the evening or on weekends is nicer.

    Also, publish-or-perish is a terrible incentive structure, this does play a part.

  • Depending on how good you are at making friends, your first conferences and workshops will be sad and lonely and exhausting.
  • Your sense of self-worth will become tied to your sense of how your research is going.
  • When your proofs or experiments are not working, you will feel frustrated and miserable for days/weeks/months on end.
  • When your theories and data do start working, the world will shine and your heart feels made of clouds and it will be blissful. For an hour or so, until you find the irrecoverable bug in your proof.
  • Academia is a scam. The fraction of your time reserved for research is a strictly decreasing function of time. Research directions are chosen because they are uncrowded and occasionally interesting, not because the outcomes would have any relevance to the world at large. Every single postdoc is sad and lonely because they have to move long distances every year or so, preventing them from having any friends.

That said, getting to do a PhD is undoubtedly a privilege. While the whole world is burning, trillions of non-human animals are brutally slaughtered every year for needless human consumption, and 815 million humans are too poor to afford enough food, you get play a part in the Unidirectional March Of Human Progress™ by spending four years in your safe, rich bubble thinking about a useless problem that Erdős once mentioned. Be sure to enjoy it.