2y ago

Sam Altman feels Silicon Valley has lost its innovation culture, saying great research hasn't happened there in a 'long time'

www.businessinsider.com Sam Altman thinks Silicon Valley has lost its culture of innovation

"Before OpenAI, what was the last really great scientific breakthrough that came out of a Silicon Valley company?" Altman said on a Wednesday podcast.

Sam Altman feels Silicon Valley has lost its innovation culture, saying great research hasn't happened there in a 'long time'::"Before OpenAI, what was the last really great scientific breakthrough that came out of a Silicon Valley company?" Altman said on a Wednesday podcast.

You're viewing a single thread.

32 comments

I would go further: the idea that great research comes out of the private sector is a myth perpetuated by self-aggrandizing corporate heads. Even most AI research is the result of decades of academic work on cognitive science coming out of universities. (The big exception is transformer technology coming out of Google.) mRNA vaccines are based on publicly funded university research too. All the tech in smartphones like GPS and wifi comes from publicly funded research. The fact is, science works best when it’s open and publicly accountable, which is why things like peer review exist. Privatized knowledge generation is at a disadvantage compared to everyone openly working together.
The private sector is very good at the consumer facing portion of innovation, like user experience, graphical interfaces, and design. But the core technologies, with rare exception, almost never came out of the Silicon Valley.
- I’d go even father - the private sector isn’t even that good at UX/UI or design
  Its main benefit is figuring out the minimum viable product and shipping it at low costs compared to the ideal perfect product from public and open design
  The private sector is way better at “we won’t spend anymore time at this. It’s good enough, just deliver the product” than the research sector
  
  isn't even that good at UX/UI or design
  Open source projects don't excel at it either, it took GNOME 20 years to stop looking frozen since the 90s. KDE is a toggle and checkmark mess.
  Only the users know how to cater themselves, AOSP derivatives, UNIX/Linux rice, seasoned designers copylefting/giving away typefaces and assets, orgs advocating and implementing accessibility options in video games, etc, etc.
  
  Open source projects have trouble getting designers to help, or developers that want to implement the designers vision
  My point was that even UI/UX research falls into the same categories mentioned by the other poster - most of the research is being done publicly and the private sector is just implementing it and selling it as cheaply as possible, same as the example of GPS
  
  As you can see from my original comment, I’m no knee-jerk defender of private sector innovation, but I don’t think I agree with this. I love open source software, but the UI is often clunky and unintuitive, like Gimp or LibreOffice. Even when it’s good, it’s often because it mimics the major commercial software.
  The heuristic I have is, when the end result benefits from communal information sharing, public is hands down better than private. We have an opioid crisis today because privatized proprietary medical research didn’t receive the same scrutiny from the scientific community as public research. Science and secrecy are incompatible.
  But when the end result benefits from a small group of opinionated people getting their way, private can sometimes be better. And good design is more like the latter.
  
  My point was that UI/UX research falls into the same categories as you mentioned. The private sector doesn’t innovate in design any more than it innovates in GPS
  Open source has issues with design more because of who contributes to it.
  If you want truly horrible UI/UX, look at tools written by hardware companies like their flashing tools or JTAG tools ;)
  
  Ah I see. Insofar as UI/UX research resembles science, and it certainly often does, I agree that it would be better if it was public not private. But as much as I dislike corporations patting themselves on the back, I just don’t think it’s realistic to say they never innovate anything ever in designing a product.
  Here’s an example: every part of the first iPhone in 2007 was already invented before its release. None of the core technology was new. But I think it’s hard to deny that Apple innovated in packaging it together in a useful attractive product.
- You hit the nail on the head. The computer mouse and the TCP/IP protocols were invented at Cal Berkeley. Stanford was doing AI before it was cool, hell before computers were even cool. Silicon Valley’s history is actually an academic one, not a private one
- In my experience, even the consumer facing portion suffers from lack of innovation. Look at the big rounds of lay offs this year, fairly uniformly one of the hardest hit teams has been UX Research. If you've worked with a good researcher, you really know their value. But translating value that into hard metrics is tough. A lot of the time CEOs in the private sector will accept Good Enough & fast Vs moving at a reasonable pace.
  I don't think there would even be an appetite for hard research at most companies. Takes too long, too much of a risk, the boss' cousin had a really cool idea instead....
  Private sector is very good at operationalizing existing technology. Outside of the FAANGs(/MAMAAs) being good enough is too easy, or investing in research is considered to be too high an expense with no guarantee on return.
  
  Yes I noticed that too about the tech layoffs. Especially nowadays, corporations seem extremely uninterested in competing to make newer better products.
  I think maybe we mean the same thing when you say “Private sector is very good at operationalizing existing technology”. The Nintendo Switch was never a technological marvel, even when it was first released. It’s an attractively assembled collection of other people’s technology. The Switch is “innovative” in a way, mostly in functional product design, but it’s not science.
- Xerox Parc brought us the gui, the copier, Ethernet, and the laser printer. Sometimes private companies in Silicon Valley actually do innovate.
  
  There is truth but also some corporate myth making here. The main elements of the modern GUI was presented during the so-called Mother of all demos by Douglas Engelbart. Engelbart would later work at Parc to make working prototypes of his ideas he already developed at the Stanford Research Institute. Now the corporate side of the history dominates the story. Same with Ethernet, which was an extension of a researcher’s dissertation work. This sort of corporate historical revisionism is exactly what I’m addressing.
  
  Damn, that was an amazing rabbit hole you sent me on.
  https://www.youtube.com/watch?v=B6rKUf9DWRI for a quick look at the demo
  
  Here is an alternative Piped link(s): https://piped.video/watch?v=B6rKUf9DWRI
  Piped is a privacy-respecting open-source alternative frontend to YouTube.
  I'm open-source, check me out at GitHub.
- The private sector can’t afford that shit. RCA famously tried
  
  The RCA story is really interesting. Technology Connections has a video on it.
  I'm not sure it's fair to say that they couldn't afford it, but they couldn't afford undirected research.
  
  Even if they hadn’t gambled it all away on Videodisc, RCA would have eventually been gutted by venture capital and then the brand sold to a cheap electronics manufacturer anyway.
- You seem to know things and I'm curious, I have it in my head that it's somewhat typical for university researchers to be able to profit off their own research by spinning up startups and whatnot. My recollection is the university typically has a contractual right to a cut of that. I don't remember if that is common, or just something in the drug research kind of sector, or if my understanding of research is based more on the Ghostbusters movie.
  I agree with everything you said though.
  
  Typically, universities own the patents to any ideas developed using university resources, including the research work one does as a professor. The university can also give a cut to professors.

32 comments