>As for the Media Lab, they're really selling the image of technological progress.
Disclaimer: I've worked with researchers at the DCI group in the media lab on hash function cryptanalysis.
The media lab is made up of a large number of research teams doing very different stuff. It is unfair to just say "the media lab does X or operates like X". Look at the research from the DCI group at the media on zero knowledge proofs systems [0], hash-based accumulators [1], or work on Ring Signatures [2].
[0]: zk-sharks combining succinct verification and public-coin setup https://dci.mit.edu/zksharks
[1]: utreexo: a dynamic accumulator for bitcoin state https://eprint.iacr.org/2019/611.pdf
[2]: It wasn't me! Repudiability and (Un)claimability of Ring Signatures https://eprint.iacr.org/2019/135.pdf
I have no dog in the fight, but I feel like if any entity is branding itself as a discrete entity, then you take the lumps of the branding just as you take the brand recognition. If you're "part of the MIT Media lab", then this affects your reputation in the eyes of the public.
Oh for sure, the grad student get value out of it because there's someone established advocating their work. For people w/ no reputation and marketing know-how, being cross-selled like this is of value, in fact it's one of the primary value add of the PI.
> The media lab is made up of a large number of research teams doing very different stuff.
Yeah fair, wish I knew more about cryptography so I can comment. In general, I don't think narratives are "bad" per-se, or they are strictly less valuable than rigorous mathematically proven technical work. For example every nation has a shared myth/narrative right, and US's myth of freedom of thought and social equality encourage lots of people to try hard, which I think of as a good thing in general.
The most common complaints form the inside are political games or PR games... like here. Either whoever has the best connections wins or whoever has the best (most?) image wins.
Remember Banksy's shredder prank? He shredded a multi-million-$ drawing at the auction. Presumably this was some sort of angry gesture towards the art-buying world. I think it's the same sentiment Ted expresses here.
A lot of researchers are idealistic, like many artists. They don't like "hype and hokum," and they take offence at how much influence PR has. It's cheapening. The value, potential or importance of research is being determined by a spokesman who doesn't understand (or care) about his own inaccuracies. The granter is obviously dumb too, judging by the dumb PR used to court them. Neither seem to care about truth. How is it possible that these dummies decide so much.
I'm guessing Banksy hates that idiot art dealers and buyers decide the value of art.
Them's the breaks though.
Also, the story of the shredder being prepared years in advance reeks of bullshit. And if that part of the story is bullshit, I can't help but question the rest.
That said, it's still a commentary on the "art world." The unidealistic ways & reasons art is bought and how the value of art works... contrast to the idealistic artist destroying his work rather than having it denigrated. It's also inevitably hypocritical because the artist is either benefiting from the whole hype and hokum or wishes to be.
My guess is that the last part is why this is such a common navel-gazing motif in art. The heart at odds with itself. Idealists always struggle with all the hypocrisy, especially their own. Poor things.
If he wants to do some really interesting somewhat outside-the-box work, he often has to find some way to make it look incredibly dull and inside-the-box and incremental in order to convince some funding bureaucrat to fund it.
Most money is dumb and flows in a herd-like manner into things that look superficially like things that have been successful in the past. This includes stock market money, government money, and yes even VC money. You often have to trick the world into funding genuinely interesting or productive work. Without a bit of marketing money will just chase the latest bubble or fad. It's like you have to save money from itself. If truly creative and productive people don't lay on the marketing, money will instead flow to charlatans who do... see also WeWork, Juicero, etc.
The last few decades of US economic policy has been based on the idea that private investment is more effective than taxing and redistributing money. But as it turns out, it's not just the government that's dumb -- all large investors behave in this dumb, herd-like way. And at least governments are more comfortable spending money without expecting a return.
I really think the power of capitalism lies not in some magical ability of markets to make good decisions but in the freedom it affords to route around bad decisions and the bureaucracies and corporate institutions that make them.
In a vertically integrated Soviet-style economy a single bad bureaucracy results in the total failure of that industry. It's not really possible to sidestep it; there is no "permission free innovation." In a free market those big dumb slugs still exist but it's at least possible for some people to go around them.
I have a pretty hard time with the theory that Doctoral MIT graduates are an exploited underclass.
Doctoral students everywhere are overworked and underpaid. Same as almost everyone, really. But the Doctoral students do get to walk away with, and heavily monetize that credential at the end of it.
And if they are doing work on aspirational “pipe dreams” pursuing a dream of the future that doesn’t pan out? Rather than what, punting on the dream and getting back to the hard work of CRUD apps and monetizing clicks?
Ha! If only! To continue doing research, the usual next step is a postdoc (~$50k/yr). Industry positions pay better, but outside of ML, there aren't tons of them and biotech (for example, which the media lab does) salaries aren't crazy high.
I'd bet that anyone with a PhD could have made more money with a masters degree and a switch to industry.
This is common knowledge for those pursuing CS PhDs. You don't do it for the money. If you only care about money, a master's degree is what you want.
The career consequences are also weirdly mixed. Some places seem to recognize that, along with your area of specialization, getting a PhD also involves a fair amount of project management, writing, etc skills. Other places (or even different people at the same place), seem to think it's a glaring red flag that you can't "get real work done" because you sit around all day in a smoking jacket, thinking. (I think that's mostly bunk--academia moves fast these days, but that sentiment is nevertheless not uncommon).
And, if anyone has actual advice on monetizing a comp/neuro PhD, I'm all ears :-)
Second, well, private investors give money to whoever they want. If they believed they'd get most out of their money by just following some academic metric like citation count or impact factor (and throwing money at them), that's what they'd likely be doing; I don't think this would work all that well for them, do you? For one it's very hard to come up with any objective metric that can't be gamed, distorted, etc. The Media Lab has a rich history of collaboration with industry and yes developing innovative products/ideas. They promote this history, and I don't think there's anything wrong with that.
That's such a broad brushstroke.
Consider, for example, Echo Nest's research in machine listening in Barry Vercoe's Music, Mind and Machine group. This all made it into Spotify when they acquired Echo Nest. His work led and pushed other researchers in the field at ISMIR, etc. Hard to downplay the impact of Spotify on our culture. This is not "hokum."
And as I shoulder-surfed the whole thing, I can tell you it started out as a crazy dream. When Brian Whitman and Barry began it, CPU power was meager, GPUs were scarcely recognized as general computation aids. It was mostly about music similarity instead of social features. (I like to think I helped Brian think that way, egoist I am). Even storing 1,000,000 songs was once difficult. This was in the AJAX and perl days.
But like Brian's daydream, many other projects there are daydreams. That's the point of the ML, to try the daydreams other people won't. Not to disrespect others for taking on research more likely to result in tangible results. In fact, that's very sane! So if you want to critique the place according to some personally-applied standards, that's great, that's your opinion, next.
It's a pretty weird thing flex about too. Like isn't streaming music based on preferences pretty easy?
That may be true, but Negroponte was proselytizing streaming media back in the early 90s, long before anything like Spotify or Netflix were actually practical. Did all this hot air have any effect? I have no idea. But from where I was sitting, he was definitely pushing hard for something that others were not at the time.
Also, the first music recommendation service I ever heard of was started at the Media Lab. It was called Ringo, and you'd email it a list of your favorite bands, and it would then email you back recommendations for other music you might like. Ringo was implemented using some sort of statistical correlation amongst other people's lists of favorite bands.
Ringo was eventually spun off into a company called Firefly which ran a music community website that used this technology. It didn't do so well, though, and was eventually bought by Microsoft, which shut down the website about a year later.
> Like isn't streaming music based on preferences pretty easy?
Sure, but I don't think that doing it very well is. I mean they typically do an okay job, but I've yet to find one that just blows me away with its ability to find music for me that I will love.
Yes, I'm introduced to some good new music this way. But I also have to listen to a lot of dreck too.
#1 is "It must have all of the music", which is really a lawyer problem, not an engineering problem.
(#2-#11 are just variations of "No no, I mean all the music", "Does it have really old music, like X?" "What about really niche music like Y?" "Oh, oh, what about Z? I bet it doesn't have Z.")
Well, maybe that's why they generally are not so great.
On the other hand, Pandora has traditionally existed with playing music that you'll like, but didn't explicitly select, as its reason for existence. Maybe that's why it seems to do a better job at this than other streaming services.
That must be why Spotify constantly fills my once useful “discover weekly” feed with what I would describe as “Ambient Indian Buddhist Trip-Hop”. Not really what I wanted at all. Though this week it seems to be “Ambient Irish Bagpipe”, which is also not at all what I wanted.
So, no... I think Spotify has a long ways to go to learn how to give me what I want...
No, I don't think so.
Millions upon millions have been spent pouring money into recommendation engines (which is an advanced "based on preferences"). Companies like Netflix know that they can keep customers that much longer if they're good at finding the needles in the haystack for each specific customer, and they famously offered a million-dollar bounty to anyone that could improve their algorithm by even a few percentage points (and, just as famously, everyone recognized that it would be worth far more than a million dollars to them).
Pandora tried to do it by creating a "music genome" for every single song in its database, with dozens of attributes per song, thus linking disparate songs though much more than just "genre."
Nowadays it's pretty much "throw a massive database at a ML system, and pump out 'people who like these 10 songs tend to like this song'", but for a long time those systems were crappy. And, even now, its hard for those systems to find a balance between showing you something that matches what you've listened to, and playing you ten nearly-identical songs in a row.
[0] http://the.echonest.com/pressreleases/spotify-acquires-echo-...
Recommendations and turning that into useful products (like auto generated playlists, mood playlists, upcoming talent, etc) are a huge draw to Spotify and creates a high bar for new competitors (including Apple Music).
Didn't the Music Genome Project / Pandora do this half a decade before Spotify?
An example I used to use with Brian was that Bad Religion and NOFX sound similar. However, fans of one of those bands weren't often fans of the other. In this model, people have a need for an amount of that sound and a social connection, NOT a need for more and more and more of that sound.
Another diff is that Pandora used human music experts where EN used machine listening. They have some overlap, and some differences, but obv the machine listening is offline.
There were a ton of music recommendations available at the time. Every music service had them: Rhapsody, Yahoo! Music Unlimited, Last.fm, Mog, iTunes, Grooveshark... Echo Nest's weren't even that good. I don't think they had song level recommendations iirc.
I'd say it be hard to convince anyone that Spotify has had any material impact on our culture...but here we are. Music (sure); the internet (sure); smart phones connecting us wirelessly to the internet (sure); a single music streaming app with countless competitors (I have heard the name Spotify, but never used it, and yet I still stream music everyday...just don't see any impact on our culture)
The explanation made sense to me and would support your point of view, but sadly I don't know if it's actually true as I heard it from someone not actually working for Spotify.
Anyway, that's certainly the hype part that your anecdote refers to. Now how about the hokum? Did your Ivy League professor draw a computer on a cardboard box and put a kid inside there? Or did they use it for a bona fide machine learning research project?
I think those are fair questions as there's at least one story of a Media Lab project that misrepresented its technology and did not really perform the tasks which it claimed to have performed.
> I asked an Ivy League professor about hype and they said the exact opposite of your Ivy League professor
The context matters, I have heard the opposite response as well, from a professor I was working for, while he was doing just that: packaging narratives w/o regard to how well the methods actually work. The honest take came at a wine-and-cheese party from a professor I was not working for. There is a real conflict of interest here, and if people stand to gain by withholding information, they will do Just that.
It’s hard to get grants to try something really out there, especially from the NIH or NSF; DARPA is a bit more adventurous. Pilot data helps with this, but those experiments cost money too, so one approach is to make the project sound almost, but not quite, obvious in light of existing work.
At the same time, it’s also really hard to get funding to confirm something that everybody “knows”, even when the underlying data is actually not very convincing. These experiments need some hype.
VCs get a management fee and a carry. They also invest some money into the fund. Their management fee de-risks them entirely.
You can't group VCs and founders on one side and engineers on the other. It just doesn't make sense.
Arguably the most financially successful founders are marketing-led, not technology-led. And some are perfectly happy to set up low-to-zero prospect companies, raise millions through networking and confidence, spend the millions with a good chunk going on their own salary/expenses, apologise to everyone in a punchy yet apparently humble post-mortem when the idea fails, and then do it all again.
In reality startup land is a scammer's paradise on all sides. Some people knowingly run a con, some are delusional, some are sincere but unlucky and/or lacking the skills/contacts/presentation to make a good idea work.
All of the above can get funding, especially in boom times. So there's surprisingly little pressure to build an ethical, product-focussed, engineer-friendly, stable business with lasting potential, and plenty of reward for not doing so.
It shouldn't be a surprise if there are people who take the bait - on any of the possible sides.
Sounds like reputation afforded to academia (professors) might be inappropriate. Besides, science shouldn't be judged by reputation but by fact and result. The idea of evaluating the quality of an idea by one's social standing is sort of odd... for someone who should be preaching science.
As the financial sources are by definition external, the people behind them have no clue about the actual research. So what can the researchers do to close the sale, if the actual science won't work? They necessarily have to tell stories.
Maybe the problem is that the science won't work and they're trying to spin it otherwise.
As for the Media Lab, they're really selling the image of technological progress. And Corporate entities find value in this exact image because they're just reselling it to their shareholders, many of whom are dumb money. But the image is required to attract idealistic and clueless grad students who'll sacrifice their lives and sanity to make demos. I know several people who graduated from the Media Lab with masters/doctorates, and they are all of this exact mold: taken captive by pipe dreams and their own delusions of what this stuff could do, while their advisors are churning through students and packaging new narratives.