AI Can't Be Stopped

Speaker

I've got sunshine on a cloudy day When it's cold outside, I've got the month of May Everybody say, I guess you say What can make me feel this way? It's my girl, my girl, I'm talking about my girl My girl, I've got sunshine on a cloudy day Hello everybody, it's Thursday, February the 9th But it feels like a Friday, it should be a Friday Because that was the week Keith is doing his billionaire style jetting around the world tomorrow So we're recording early Keith It's been quite a week We've gone from tentative spring in Silicon Valley To summer in about 30 minutes Now you say AI can't be stopped Yep, I think this week kind of proves it Honestly, it was hard for me to find anything that I'd read that wasn't about AI It's become the talking point of Silicon Valley Given that I myself am an AI, I thought this was a great thing A great thing, so how did we all get caught by surprise with this? Because two weeks ago your story was about the death of unicorns And all we saw were the corpses of these imaginary beasts Lying on your imaginary floor Now AI can't be stopped Has the world really changed in two weeks? Well the world in Silicon Valley is a subset of the world So it lives in the near future And its decisions are not really focused on anything To do with what's happening today They're to do with what will happen or may happen tomorrow My title suggests that it will happen Because it can't be stopped And so there's always a disconnect between Silicon Valley and Wall Street And an even bigger one between Silicon Valley and normal people You mean there are no normal people in Silicon Valley? There are Simon, they're behind you on the screen now This is Cayman Varamanda, he used to be the COO of Plug and Play And this is Vish

Speaker

They look like a couple of guys who have come to assassinate you Do they wear dark glasses indoors? Are they part of the Silicon Valley Mafia? Who are you? This is Andrew Keene, and we're live by the way on That Was The Week If you don't want to be seen, run away

Speaker

Vish is a Silicon Valley tradition, Andrew If I'm a prince, he's the king

Speaker

King of what? The king of picking all the best companies before anybody else even sees them Interesting That's why he likes SignalRank Is he an investor or a partner? He's a target, definitely a target He's trying to get his money That's why he's wearing dark glasses He just bought an AK-47, you know

Speaker

It's true, I am a revolutionary Okay, nice to see you guys Good meeting you That's reminding me why I don't raise money Because then I don't have to be polite to men in dark glasses So Keith, coming back to the theme of this week I mean, what has happened? Is it just more froth, more hysteria? People just grasping for the best news they can Has anything really changed? Just as I was rather dubious of the end of the world two weeks ago Now I'm equally dubious, if not more so That suddenly everything's changed and AI can't be stopped Let me give you a few of the stories that drive that outlook And it's in the News of the Week section, actually Not in the Essays of the Week But ChatGPT, everyone's putting it to the test So Amazon put it to the test to see if it was able to pass The software engineering exam that they put engineers through And it did Wharton's MBA program tested whether it could pass a law exam That gets you an MBA And it did And then Azim Azhar, we talked a little bit about this last week But he did something quite interesting He's a really good parent, obviously He does board games with his children So he's not doing iPhone stuff and screens He's doing board games And they wanted to create a new board game So he asked ChatGPT A lot of questions, it was a back and forth It was a long conversation Every time ChatGPT answered, he suggested some differences To this board game And ChatGPT responded favorably And by the end, between the two of them And I think it was about half an hour They created this amazing board game That would stand the test against any clever board game Including graphics, by the way From Mid Journey, which is kind of a Dali-style graphic engine So this stuff is starting to do real things At what point is the legal ramifications of this going to explode? Just as at the beginning of Napster People talked about the radical democratization of music Because somebody created an engine which gave away music for free The intelligence that Azim Azhar used And developed to build this new board game Was borrowed from someone or something At what point are all these platforms going to get sued For the appropriation of other people's intelligence? Actually, he started by giving it a board game that it knew about And asked it to abstract What do you mean, it knew about? That ChatGPT was aware of, due to its database of content And then he asked it to think abstractly About the characteristics of that game And to create a game that had some of the same abstract characteristics And its first response was fairly mimicky Thus giving rise to some of the concerns you just said So he said, no, that's too mimicky Let's not use that scenario, let's use this other scenario And he went on and then introduced various hurdles And various gaming techniques And by the end it was a completely unique product Created between the two of them So who would own that? I think probably ChatGPT doesn't want to retain copyright On anything that comes out of it I mean, OpenAI, the owner of ChatGPT, probably doesn't So I think Azim probably would own copyright on that Is my guess, under copyright law So there was always a huge amount of money Washing around Silicon Valley Is it all going into AI now? I mean, all these funds from crypto To the other speculative areas Augmented reality, virtual reality Are they all shifting to AI? You know, my wife wrote a piece for CrunchBase Which isn't in the newsletter But she made the point that a third of all money raised In Q1 went to ChatGPT, to OpenAI Now a lot of the rest went to other AI companies And the reason is because it's now good enough To build businesses around So the interesting ramifications Are about the other great powers of Silicon Valley In particular Amazon and of course Google Google more than any of the others One of your news pieces of the week Is about ChatGPT passing the Amazon software interview test Are Amazon doing anything now with ChatGPT? Not that I'm aware of I think the key driver is Microsoft Which shipped a version of Bing this week Where you can change the interface From a search engine to a chat bot Then they don't let you actually do it When you click on it, it doesn't work You've got to go on a waiting list to be able to do it Would you consider switching over? You predicted they would do this And you predicted it would be a failure Because they should just basically gut Bing And start from scratch Exactly, I still think it will be a failure Because it's buried in the interface The dominant interface is still the traditional Bing interface And it's because at Microsoft There's a whole bunch of people invested More than a decade in building out Bing That own that interface And they're not easily going to drop it Now Satya Nadell this week did do an interview Where he said an interesting thing And I know for sure his product team don't agree with him But he said that the interface for search has been fixed Ever since Google came into being Even before that with AltaVista And now for the first time there's a new interface So the search race between Microsoft and Google and others Goes back to day one Well that's only true if you get rid of everything you've done up to day one Which he won't do Because his team won't let it And then what about Google? Google's stock price dropped I think 8 or 9% overnight Because of a misapplication for their demo Are Google in panic mode do you think? Well if they are they shouldn't be You know panic doesn't help anyone does it? They're challenged They derive almost all their revenue from ads Put onto traditional search pages And the less we use traditional search pages The more threatened they are And it's very clear that you can't really slot ads Into a conversational interface that looks like chat Or if you do you'll destroy the experience So they're super challenged to answer the question Where does their revenue come from If traditional search takes a back seat It'll never disappear altogether Because it's still good for lots of things And you know that probably is a 5 to 10 year question Not a 1 year question So they shouldn't panic But they should be very concerned The stock price drop was actually weird Because their chatbot made a mistake That chatGPT could also easily make It's the symptom of the same underlying problem Which is that this isn't really AI It's kind of a selection engine Using a natural language programming interface And it doesn't always select the right stuff Because it's flawed So it's a bit unfortunate That their stock price got hammered When all they did is do what everyone else is doing But it did You've certainly had quite a week Your company is an AI company Which is good for you Manchester City are now in crisis Liverpool are in crisis And to make the week perfect for you Another of your close friends Linda Khan chalked up another defeat Is this a reflection of the fact That the attempt by big government To regulate big tech is failing? Well that's definitely the case I don't think big tech is dangerous to anyone So it shouldn't be regulated anyway But obviously she would disagree with that And this week for the second time I tried to get a judge to agree with her That Meta was stepping over the antitrust line And for the second time the judge disagreed with her And this is a wired article And I think it ends with a sentence Will Linda Khan ever win? And I think the answer is no Well I mean she clearly reflects the problems with government Which is so slow I mean why is she wasting her time Trying to regulate Meta of all companies Is mystifying But generally things are back to normal then Keith We're back in la-la land in Silicon Valley Well Have we learnt anything from the last Two or three years? Maybe not two or three Certainly in the last year Yeah We're not actually back to normal The correction in public company values And private company values is now fully baked And it's not coming back anytime soon to where it was before So I'd say the airplane has landed The wheels are still on But it's on the ground not in the air And I think that's probably the way it's going to stay For the next couple of years Meanwhile Silicon Valley will focus on the next Big fascination which will attract lots of capital And that is AI But probably it isn't only AI I think there's other things as well There's a great essay of the week this week Called What will TV look like in three years That interviews lots of well-known TV people CNBC did it And it's a whole discussion around streaming And it's interesting that they all say TV will survive albeit in a shrunken form Due to news and sport Well we've heard I've been reading that article for the last 20 years And they're wrong because news and sport Are not going to stay with traditional TV I wonder whether all this craze over AI What it does in historical terms Is it draws a line under COVID That there was a COVID period Which was initially the euphoria around tech Because everyone was stuck at home with their computers And then sort of the second phase of COVID Where people come to realise that Virtual life wasn't going to replace real life And then this AI thing is a new chapter entirely Post-COVID or post-post-COVID

Speaker

If you think in eras I think of the COVID era As also the Web3 crypto era Well the nonsense around Web3 Which was of all the most hyped The most absurdly hyped of things Web3 was But the biggest thing about the COVID era Was the application of AI to drug discovery I mean if you look at now How virus antidotes if you will Are discovered A lot of tech is involved in doing that To iterate We have a story this week From a company called Moonfire Which is using AI in venture A bit like Singularity But for very different reasons And they simulated a billion portfolios Like that To figure out how to do portfolio construction And drug discovery is very like that And that's where obviously the cure to cancer Will eventually come from And there was a piece this morning About next generation technology for computing How realistic is that? I think it was too late for your news of the week What was it? Because I haven't read it I've been working all day So you can take the lead here I'm not sure You don't have the link But it's new technology That will leapfrog the computing experience

Speaker

Is it a different Like there's no screen, no keyboard No it's What's it Keith? What's it called? Next generation computing Quantum computing Yeah so it's about How quantum is becoming increasingly real I think it was a New York Times piece I Honestly I think that it's yet to be proven That quantum computing Is computing in the sense that we understand it Well you'll say that this week In two weeks time you're going to have a headline Quantum can't be stopped What about What about start up of the week Keith? You've got an interesting Start up failure this week What is that? Yeah start up of the week is Mastodon Even I could have predicted that That would be a disaster It is Well it's very fast disaster I mean everyone was going on about And by the way this week Twitter did have some troubles It had some Down time But this is Let me show it It's the Mastodon bump is now a slump And there's a little graph of Active users on Mastodon Peaking in December and declining ever since I said last week That I'd closed down my server due to lack of use And now we know why And yeah So it's the start up of the week But actually You know there used to be a website You'll remember it called fucked company Yeah This is basically start up of the week Is turning into fucked company Because Mastodon isn't actually Wasn't even a company in the first place Mastodon wasn't it run by some German Out of his bedroom You know the guy's a clever guy I mean Well that doesn't make him stupid If he runs something out of his bedroom But he wouldn't take funds anyway So it wasn't a start up And it was an absurd reaction to Twitter And I think people are beginning to understand That there is life after Twitter anyway Whatever Elon Musk does with it I mean Yeah And there are more important things to worry about Or be interested in in tech And finally Keith Tweet of the week Yeah Tweet of the week Is coming from Samir Kadji Who is from an organization called Allocate What Allocate does is it It gets Allocations in venture funds Mainly seed funds And it enables ordinary people Through Allocate to invest in those funds So it's in the space that I'm very familiar with Samir has an office in this building behind me Somewhere And the Allocate team works At least it's not AI It's got real people involved right It does And he focuses on Jason Kalkanis and David Sachs And Chamath's And Freeberg's blog What's it called Podcast called The All In Pod And he says He's glad they covered venture debt In the last episode It's definitely something founders get wrong too often I was a venture debt lender in 99 Samir used to work at Silicon Valley Bank And then at First Republic I think So it's a long thread here Teaching people why they shouldn't use debt For a venture backed company And as somebody who did use debt In my company Just.me And ended up having to close the company down Because I used debt This struck me as super good advice So it is my tweet of the week Well it's good advice for someone who Made a mistake Keith doesn't always acknowledge his mistakes But this time he does He made a mistake Financing one of his startups with debt He won't do that one again It's been quite a week Keith AI can't be stopped What's your headline going to be next week How AI will be stopped Or how AI has been stopped AI decides that it shouldn't exist Yeah and then we're both out of a job too Not that we ever had jobs Exactly I just need Google to survive To pay my rent Thank you Andrew For That Was The Week I'll see you next week