There's an article like this once a week because everyone is trying to stake their claim to "I predicted it" so that in 10-15 years they can retire on the conference circuit (ala yann lecun and bengio and all those others).
Here's the truth: the future of AI is commitment completely unknown and not even at all certain - plenty of revolutionary technologies have fallen to wayside over the years.
So to my fairly-educated-in-this-area mind I think we should stop listening to wannabe thought leaders (and even minted thought leaders) and just keep focused on our small parts.
This take is both overly cynical and downplays the contributions people like Lecun and Bengio have made to the field of machine learning, both through their own work and their academic progeny. I see you work broadly in the area, but I doubt your research touches on anything they've affected (that you know of) if you earnestly feel that way.
This article's focus doesn't strike me as especially aligned with current problems in applied AI (how will self-organizing systems relate to prediction problems in NLP, tabular data, or computer vision?) but the connections to robotics are plausible, and in any case the tone doesn't come off as a wannabe thought leader trying to stake a claim, more like an excited learner who is really interested in some new ideas in a niche that may end up being narrower than they currently hope. I won't be assembling object detectors this way any time soon, but it was still a really pleasant read.
>10-100us for simple models (say, a fully connected dnn with a few million parameters)
I basically don't believe you. I'm a researcher in this area (DNNs on FPGAs) and you cannot get these latencies on real models without going to FPGA (and you're not synthesizing Verilog from TF, unless you're one of my competitors...). Just your kernel launch overheads for GPU are on the order of 10ms. For example, here's a talk given at GTC a couple of years ago where they do get down to 35us (on tensorcores) using persistent kernels, but on a mickey mouse network
CPU (where you don't have to deal with async CUDA calls) won't save you either; again here's a paper from USENIX (so you know it's legit) that shows that lowest times for real networks on CPU are ~2ms (and that's on resnet18, far shy of "millions" of weights)
Sorry let me restate this in a way that seems less targeting.
There are many areas of disconnect with people in general in what they think they want versus what they actually want.
For example: Money. Many people (men especially) think they want lots of money. They don't. What they want is happiness. There's even data to back up the fact that past a certain point money doesn't make anyone happier. Yet the drive to become successful is so strong that men will sacrifice their happiness for a shot at more money. Why does burnout even exist? Because men desire money so much that they illogically sacrifice their happiness for something that won't bring them any more happiness.
The above isn't absolutely true but it's a generality that people understand to be generally true.
Like men and money... women and dating is one of these psychological disconnects humans have.
The problem with many people and with you is that everyone has all these psychological triggers that are always on the look out for anything sexist or racist. Relax man, not many people are racist or sexist and HN should foster an environment where people are able to talk about real biological gender differences without getting cancelled.
No I expect mature discussion as the rules of HN specify. There are plenty of people with opinions I vehemently disagree with. The death penalty for example. But their opinion is their right and instead of attacking them and calling them murderers I openly listen to what they have to say and address their points with my own.
Relax. My opinions don't even approach the extremes this paradox is referring to. It's akin to a woman saying all men are only interested in Sex. Probably not true for many men, but also very true for an equal amount of men. Either way, not a big deal.
There's no holocaust denial here or anything like that. Also women and dating advice isn't even some kind of controversial hard hitting topic that effects anything.
We can have mature discussion about the merits and disgusting parts of a death penalty which directly influences human lives but we can't do so about women dishing out bad dating advice? Come on man.
I beg to differ on the happiness part. There's a bit of a high when they win... and it's fleeting... but in general the money trophy does not lead to happiness IMO.
>"then why are you going through all the rejections with the associated pain which makes me suffer so bad?"
you've assumed the premise without even knowing it and it's doubly ironic because it didn't occur to you exactly because of what you represent in your first comment.
here's the answer that didn't occur to you because you just assumed you were right: lol what rejections and terrible suffering and pain?
i'm recently married but during all of the time i dated (>15 years since high school) i had maybe one or two rejections at all (and none to speak of that were painful). and i had 2 relationships a year over that time (on average - college was college...).
you know why? well for one because i did exactly what this person you responded to suggested: never objectified women and was always confident about my merits. the other thing i never did was treat dating like some kind of obsession or game, so never did i "swing for the fences".
>Cam Newton
coincidentally enough i went to college at UF and my time overlapped with Cam Newton's. if he's the guy that is a symbol of success for you (even metaphorically) in this domain then ho boy you have bigger problems than just dating.
Seems to me a failure to approach at the bar. After an intial approach (which could also be rejected outright) maybe 10 minutes will pass before the moment of truth comes and you either get the number or not see each other again. At least in big cities like NYC, LA, Chicago..that's the most frequent scenario as far as nightlife is concerned.
Not getting the number=rejection
> if he's the guy that is a symbol of success for you (even metaphorically) in this domain then ho boy you have bigger problems than just dating.
Well, luck plays a huge role in every person life..that said if you think Cam Newton is a failure where are your 100M dollar for "working" in the NFL playing football?
And of course I imagine he had so many partners he lost count (and also being sexualized by a good % of the millions of women follower he has on INstagram)
Again lots of it is luck, but it cannot be said that Cam Newton is not lving his best life according to objective metrics.
you're redefining success to be a very narrow, highly, contrived, outcome. i had a rich and successful dating life by any measure so you can't tell me that i wasn't successful (you're just arguing that water isn't it if you do that).
>Makes me wonder how much faith I should put in comments on topics where I don't know as much.
Despite murmurings to the contrary hn is no different from any other social media site that doesn't curate content. The fact that the denizens are techies or whatever doesn't supervene that fundamental fact. So the answer here is you should trust comments on hn just as much as you trust comments on Facebook.
As another funded academic in an area that comes up very frequently here (deep learning), personally I use hn as a link aggregator for interesting content and pay very little attention to the discussion.
>They had a language requirement for a math PhD. To get a PhD you had to either be fluent in one of French, German, or Russian, or you had to have a sufficient reading ability in two of those to read the current and historical math literature in those languages.
This a thing of the past, from a time when people in the USSR still published things of note in Russian (the same for the other countries/languages your mentioned). Today everyone publishes in English and I don't of a single math department that still has that as a requirement.
I also don't know of any other such extracurricular requirements for any program (can't even imagine what it would be - community service???)
Hmm must be the last (leave it to Harvard to persist in the persist in the pretence I guess) but fwiw this is the entirety of the task:
>Students may request to substitute the Italian language exam if it is relevant to their area of mathematics
Which is far far short of "learn a language". But these requirements were always like this (a pretentious formality completely insignificant in comparison to the actual research and course work).
Lol this is so aspirational it could only come from an undergrad.
Let me tell you that I've finally made it to the stressful part of the being a serious "AI" researcher, where I have a real project (as in difficult to achieve goals, not just "turn the crank" stuff) and real deadlines (deliverables on collaborators projects and my own conferences submissions) and the only thing I prioritize above doing the work itself is keeping my advisor (and other collaborators) up to date on what I'm doing so that when he reads my paper draft he's not completely lost. Everything like organizing papers, citations, logging infra, etc is meaningless when you're trying to piece together a solution. Like seriously somedays I barely have time to exercise and eat dinner with my wife (let alone organizing my bookmarks).
For example I'm trying to solve a particular compilers problem using integer programming (note that at a high this isn't that high level because this is a small cottage industry) and so I have like 50 paper tabs open that I bounce between when thinking/experimenting. The way it usually goes is I'll hack, get stuck, go back to the papers, find something, hack, and on. And usually the eureka moment comes some hours later because I connect something.
You might say that I'm a bad researcher but I know for a fact (external validation) that I'm not. And if you look at other highly productive researchers (like TT track profs at my "elite" school) this is indeed how they work. All of this zotero, notion, mlflow stuff is of the ilk of productivity porn for other flavors of knowledge workers (ie a mirage and/or snake oil). Let me put it this way: my advisor is a top 500 h-index person (the exact significance of that metric notwithstanding) and he doesn't have a bibtex of his own papers, let alone zotero for all of the papers he reads/comes across.
The only thing that matters is code/math/etc output (whatever your material output is) and your abilities are also highly correlated with it with the casualty flowing in t opposite direction (make more stuff and you'll get better at making stuff).
But I guess conversely do do some of these things when you're young and have the time (and I don't mean that condescendingly). E.g. reading outside of your area is probably the most valuable (from my own, admittedly a typical, experience, since I jumped domains many times); I very frequently can outpace even my senior colaborators very quickly on understanding a problem and solution simply because when I was younger I dabbled in ... all the things (physics, math, cs).
The other thing that I'll say is there's something obviously missing from this list but only if you've really made it this far: collaborators and interactions with collaborators. The only thing that matters aside from the produce is getting people to make use of it. That means writing, speaking, and getting buyin from your collaborators. If you really truly want to be successful then work on your people skills as it pertains to this area - that means learn to speak the language of your research community, learn to give good (engaging, interesting, useful) presentations, learn to write well (including making nice diagrams), and learn to explain things in ways that smart but busy people will understand. Besides all of this being key to being productive it's also what feeds you (i.e. the real #1 priority) since it gets you jobs, academic and industry.
"but", Zotero is really neat if you use the plugins which make it just one-click-to-save it in your Zotero from browser. Hardly any overhead with that.
Here's the truth: the future of AI is commitment completely unknown and not even at all certain - plenty of revolutionary technologies have fallen to wayside over the years.
So to my fairly-educated-in-this-area mind I think we should stop listening to wannabe thought leaders (and even minted thought leaders) and just keep focused on our small parts.