HN Explorer

Ask HN: Do you know who this person is?

22 points by dxtr ↗ HN
Hi! Quite a while ago I saw a tech talk by someone.

I'd like to see this talk again.

The problem is that I can't remember anything about the person in question and I can't remember the title of the video.

All I remember is that the talk was about being quite cautious about using technology we don't understand (Quite a lot of emphasis was on machine learning) and the speaker was actually quite funny.

The talk was on one of these coding conferences. Don't remember which one, obviously.

The speaker in question has a homepage on a university website (I think it was a university in the US north-east but I'm not sure) and was actually really funny. He also had other talks he had made.

I've been tearing my hair out here - even spending a few hours scrolling through my Youtube history.

Do you know who I'm talking about?

11 comments

[ 2.8 ms ] story [ 29.8 ms ] thread
Tomte
James Mickens, probably: https://www.usenix.org/conference/usenixsecurity18/presentat...

Many, many people found that funny. I don't, but I seem to be in a small minority.

His Mossad article is great, though!

andyidsinga
hmmm - what happened here is very interesting outside of the specific question : HN as human based search with complex criteria. Looks like OP got his answer in ~1 hour.

I wish I knew how many people viewed this question during that hour.

alphor
Yep! Amazing how all we need now to get a flood of info is a keyword.
S4M
FWIW, I already asked similar stuff ("What is the movie where...?", "what is the book where...?") on stackexchange and I could get the answer relatively quickly.
shoo
> I wish I knew how many people viewed this question during that hour.

here's my naive attempt to crudely estimate this:

http://www.siteworthtraffic.com/report/ycombinator.com

  > [daily] Unique Visitors 	257,211
  > [daily] Unique Pageviews 	439,687
https://www.similarweb.com/website/news.ycombinator.com

  > Total Visits 15.88M  [per month]
  > Pages per Visit 2.51
https://ahrefs.com/blog/website-traffic/

  > SimilarWeb overestimated total unique visitors for 91.67% of the websites.
  > SimilarWeb overestimated total unique visitors by 308.81% on average;
  > SimilarWeb overestimated total pageviews by 210.58% on average.
SimilarWeb estimates 1.3 million total daily pageviews . Lets blindly correct trusting the ahrefs comment about overestimation, that gives us 1.3 / (1 + 2.106) = 0.42 million total daily pageviews. That roughly agrees with the siteworthtraffic 0.4 million unique daily pageviews number. So, let's arbitrarily trust siteworthtraffics 1/4 million unique daily visitors estimate.

Let's assume that traffic is uniformly distributed over each 24 hour period (wrong) and that 50% of traffic hits the first page of links under /news, /ask and /show (wrong) and that 50% of traffic hits other links. Of the former 50% of traffic, lets assume that 50% of people who click the link also scan the comments (wrong), and that views are proportional to the number of votes each link has (wrong), and that the current selection of links and votes is the same as during the 1 hour period between the question being asked and answered (wrong).

the sum of votes over links on the first three pages of /news, /ask and /show is roughly 4200 votes, and this ask topic has 19 votes, giving it a weight of 0.0044 .

multiplication gives 250,000 unique visitors / day * 1/24 * 0.5 * 0.5 * 0.0044 = 11 unique visitors during the 1 hour period before the question was answered.

it wouldn't be surprising if this estimate is off by 1 or 2 orders of magnitude -- underestimating, not overestimating, obviously -- if you can propose a more plausible estimate, propose away!

matt_the_bass
Nice analysis. No idea if it’s correct but it seems logical.

Ps this would be a great question for an interview (if you believe in those “how does this person think” type questions).

Tomte
Write hn@ycombinator.com. Maybe they are in the mood to look it up.