Ten Problems for Artificial Intelligence in the 2020s | Hacker News Reader

Ten Problems for Artificial Intelligence in the 2020s

tenproblems.com

10 pointsDrNuke8 comments

This is just a spam submission that keeps being posted by the same user to flog their booklet site.

@mellosouls DrNuke == Giuseppe Cornacchia, the author of this content [1]. That's why he is so defensive and has a vested interest in seeing it hit the front page.

The format seems to be to pick a topic, spend ~10 days researching it and then write a booklet preaching about its future. Apparently the goal is to "provide an informed report based upon peer-reviewed documents, hopefully debunking fake news, sensationalism and outright biased agendas" [1].

@DrNuke No offense intended, but some people spend their entire lives researching these subjects and fail to ask even the right questions, let alone touch the right subjects. The themes you have pulled out for AI for example seem to be generated by looking for common themes in the last year or so, which has a major bias towards research that can be conducted due to lockdown restrictions.

[1] https://www.tenproblems.com/about/

Nah, I do not need exposure, it is just another side project. You may make a better argument with saying that the bar of that project is set too low (academically, technically, commercially, whatever), which I can see as well and agree. Then, I also have to turn off this device, go out of my cubicle and meet real people from any environment and segment of life. That’s where the need for compiling such admittedly bland summaries was felt the most, at least becoming able to offer a few, valuable references to the fellows asking for them.

My initial comment was harsher than I meant it to be, for that I apologize.

> You may make a better argument with saying that the bar of that project is set too low (academically, technically, commercially, whatever), which I can see as well and agree.

Don't get me wrong, I think it's a great start and there is value there. The idea itself is good, the website is good, but the depth is missing (at least for AI).

I would suggest to start in the topic that you are most comfortable with and to really explore it. Bring in experts, hassle academics, push hard to really understand it.

Genuinely, best of luck with the project. Publishing is really tough.

Ok, thank you, much appreciated. The selection of the ten most relevant problems does not come from my own judgement, though: it always starts from the higher level, general references linked in the free intros. For AI, you have: Stanford Uni, the United Nations, Nature, SN Computer Science and the Montreal.ai cheatsheet. I am not sure you can get better or more representative of the current times than that.

Calm down, warrior. It links to a free intro with five references to very good, up-to-date, freely downloadable internet references for people willing a general summary of the field. No cookies, no ads, no tracing, no purchase, no subscription. Free lunch.

It's a stub article (references or no) with a chapter list of a thin booklet you have to pay for by clicking the Amazon link.

ie. It's an advert - the same as your previous submissions of your site.

Which hundreds of people, admittedly not know-it-all like you, benefit from, without forking out a single penny.

Reply on news.ycombinator.com