Hacker News new | past | comments | ask | show | jobs | submit login
Google's top quantum scientist explains why he resigned (forbes.com/sites/moorinsights)
225 points by jonbaer on May 1, 2020 | hide | past | favorite | 139 comments



I think that this article points to a general challenge with our education system. By and large, the people who stick with training all the way through a PhD and postdoc in a complex field are not hoping about being a cog in someone else's machine. In academia, I've observed that the sign that that a trainee is ready to move on to independence is that they begin to have disagreements with their mentor that don't reflect misunderstandings of fact, but rather proper, interesting diverging hypotheses. In this story, the wiring incident reminding me of this.

The problem is that solving these sorts of hard problems require both leaders and workers. In this case, and many others, there is not an educational path for training people to be technicians with equivalent knowledge as a PhD (which is why they poached an academic lab to do this). Moreover, if there was a path, history has shown that the people that accept the follower role (rather than wanting intellectual leadership) are highly dispensable. (See, e.g., the optics industry moving laser fab to Malaysia.)

It's also a problem in medicine - physicians resist loss of diagnostic or treatment control. Though in medicine, there are NPs and PAs, but they're generally treated with much less respect.

I think it's interesting that he talks about how "he" Martinis was doing the wiring. Was he? I suspect he had a team working very hard with substantial responsibility, but...


> I think that this article points to a general challenge with our education system. By and large, the people who stick with training all the way through a PhD and postdoc in a complex field are not hoping about being a cog in someone else's machine.

That's not limited to academia. In the industry (and even outside of tech) people who will have a sufficient understanding of how something works will at some point in time no longer purely want to follow orders, but want to drive things and innovate on their own.

In this situations both sides - the current leaders and someone who wants more responsibility - need make some compromises in order to try to accomodate this interests.


Thank you for writing this!


The pattern seems familiar, take a succesful, self-propelled creative inventor, typically an engineer kind of person, put that person in a group where they need to submit to ideas of others, and things eventually go wrong.

It's not that unusual to see a founder get excluded even if that person was arguably the primary driver, the deteriorating relations don't allow the situation to continue and the alternative is blowing up the organization.


This is one reason you should always just execute - never ask for permission. Have conviction - think your objective through - plan - cross your t's and dot your i's - and then just execute and show your results to everyone in your mgmt chain.

...to the degree that you can do that without getting fired.

It's always easier to ask for forgiveness than permission. Trying to convince a bureaucracy of anything is futile.


It's personality dependent. As in your actions have to align with your personality - https://en.wikipedia.org/wiki/Revised_NEO_Personality_Invent...

Some people can. Some people cant. There are no general rules that work for everyone.


like you can do that on a quantum computer machine. not everyone works on software


This just as easily applies to artists, its definitely not inherent to engineering.


In a way, engineers are a type of artist.


Yes, but management wants them to be assembly-line workers. A lot of organizational tension in tech companies can be traced back to botched attempts to force-fit artistic pegs into assembly-line holes.


Honestly, this is probably a big part of the reason I've really come to dislike things like Scrum and what Agile looks like now. My experience has been that these methodologies tend to treat devs as fungible cogs which fits very much with the idea of engineers/developers as assembly line workers.


Exactly this. Management wants every part of the production process to be easily replaceable, including the people. This is why management loves, e.g., Java. Java gives you a lot of the benefits of Lisp (GC, objects, exceptions) but without any of the features that facilitate writing beautiful elegant code. So Java code is ugly and festooned with boilerplate and using it is an incredibly painful experience (for a Lisp programmer). But if I'm a manager and I have a Java programmer working for me and one day I don't like the cut of his jib, I can fire him and instantly replace him with another Java programmer, because they are a dime a dozen and one is as good as another. As a manager, that one feature dominates all the cruft and the horror, because I don't have to deal with the cruft, I hire you to deal with it for me. And your replacability gives me power over you. So you type boilerplate while I play golf.

It would be a pretty good deal if I enjoyed golf as much as I enjoy writing Lisp code.


I'd always had the impression that this was the case, but when I started to program in Erlang after several years of Java it became clear to me how true this is. Java has been adding features that improve expressiveness, and from the perspective of a strictly Java programmer they are great. From the outside though, it feels like these new features still take way more work than should be necessary.

I appreciate what Java was meant to be (a way to drag C/C++ programmers closer to Lisp) and I would say that's a goal that has been achieved. I just wish it were easier to go the rest of the way.


Indeed, a power struggle. Management's natural responsibility is relationships to other humans, which makes them much harder to replace compared to a cog that performs a technical task, no matter how much skill that task requires.


How's the state of Lisp in production these days?


My one-man company uses a bespoke e-commerce system written in Lisp. It uses Stripe and Easypost on the back end and gives me a completely seamless fulfillment process that lets me ship anywhere in the world with essentially zero effort. I also consult for a major chip developer to help them maintain an internal design tool written in Common Lisp which they consider to be a major source of their competitive advantage. So it's not widely used, but where it is used it's a big win.


Sounds like lisp is good for a team of size 1, but are there any larger teams using lisp successfully?

I've heard theories that lisp teams don't scale well because the language is so customisable that team members have trouble reading each other's code.


That depends on what counts as a team. I use a lot of other people's code in my own projects. So although the people who wrote much of the code that I use didn't work for me, it's not exactly a solo effort either.

But yes, managing Lispers can be a challenge. That's one of the reasons Java is used so much more.


Hell I have trouble reading my own code. Dense, arcane, not descriptive of what's going on. FWIW


I love my code. Whether it's 6 months old or 20 years, I can readily understand it, and prefer it other people's code by a wide margin.


Erratum: It's too late to edit the comment, but what I was trying to say is that my code doesn't change because I don't learn. Alas, my wording somehow doesn't convey that at all.

Credits to the clairvoyance of JoeAltmeier for the kind correction!


Sorry didn't mean to be catty. Just that, our 20-year old code for most people seems naïve and clumsy.


Must not be learning? My code changes because I learn and change.


Someone with no actual experience in a team environment using Lisp just made that up, wrote it up into a glib article and that turned into a famous meme that gets trotted out about semi-monthly on HN.


What do you think of Stripe's docs? As an Elixir dev I'm somewhat dismayed at their trend towards assuming everyone is using one of their blessed handful of languages.

I mean, it's great that they've made a few official integration libraries, but if all the docs assume you're using one of them then the docs leave out crucial information about how to work with their systems for everyone else.

I suspect the information is all there somewhere but at least since SCA, it's buried. At least for me, the dev experience using Stripe has deteriorated quite a bit in the past 2 years.


Stripe's documentation is excellent. All you really need is a REST API.


There's definitely more to it than that. I wrote a very lengthy comment about my experience working through their payment intents docs with basic REST requests and hit quite a few errors due to requirements not mentioned either here or any linked page: https://stripe.com/docs/payments/accept-a-payment

It's worthy of a whole blog post, though, so I'll share it that way.


What can I say? I had no problems. But maybe you were trying to do something more complicated than I was.


Ok, here's the short version:

My experience going through the payment intents was that sending POSTs exactly as described was an immediate error for not sending a version in the request header. Nowhere in the doc was this requirement mentioned: https://stripe.com/docs/payments/accept-a-payment

The problem is step 1: "Use our official libraries for access to the Stripe API from your application"

Similarly the link on this page regarding webhooks leads to another page with similarly insufficient docs: https://stripe.com/docs/payments/handling-payment-events#bui...

If you're not using their library, there's no info about how to parse the webhook signatures and verify them. There isn't on this page either: https://stripe.com/docs/api/webhook_endpoints/

It ended up taking me half an hour to find a page that actually explained (with no examples) how to actually do it.

"Use the official library (and we have no plans to support more languages)" is not excellent documentation. TBH, it was one of the more frustrating experiences I've had with anyone's API docs in the past 2 years.


Again, what can I say? I wrote my stripe code over three years ago so I don't recall many of the details. All I know is that I don't recall having any trouble with it.


FWIW, I finally found what I needed after finding some of their older docs that predate SCA.


I think the problem lies with who owns the process. If engineers own the process, then things can work. If non-engineers own the process, then you get the assembly-line mentality. Whether one calls it "Agile" is really irrelevant.


A lot of application software is fungible, boring, and just needs to get done. Agile helps with that.

No one is suggesting planning poker for qubit wiring.


"I myself am a master of the software arts - or an engineer if you want to be a dick about it"

https://www.youtube.com/watch?v=F4MCXotzJfA


"developer" if you're a canadian


What an interesting and apparently frank interview. Martinis was the UCSB professor whose lab setup was basically purchased by Google. Apparently he and Hartmut Neven, another somewhat legendary figure, had “creative differences” and Martinis has left Google. But they also talk about Google’s big plans for quantum. Thanks for posting this!


I almost joined his lab around 2014 as a PostDoc (without knowing about the Google deal, which was not yet official I think) but eventually I stopped the application process because I wanted to pursue my own startup instead. Today I kind of regret not having stayed in quantum computing as my startup was only modestly successful and as I would probably have been much better off working for Google, in retrospect. I'm working on my next startup now but I keep wondering if I should go back to quantum computing eventually, I'm not sure though if I'm still relevant as my PhD dates back to 2013 already (I had good results though and did it in a renowed QC group on superconducting qubits).


As someone who started off in "quantum" in similar conditions around the same time, but decided to stay on the academic researcher track:

A great deal of the recent progress has been on honing, parallelizing, and automating the same processes one used to do by hand/at small scale in 2013. You'll probably instantly recognize most of the things going on in the lab, and could become productive on a short schedule.

However, I'd be wary of the medium-term prospects of the field: The discord between the perceived and actual capabilities of the hardware remains as big as ever. You could consider re-entering the field if you're happy contributing to incremental technological advances relevant to other researchers. Progress with meaningful societal or economic impact is probably much further off.


Thanks for sharing your view, I appreciate it! The point you mention was a large factor in my decision. Most of my former colleagues and advisors also "abandoned" superconducting quantum computing later to work on more fundamental research again (e.g. inventing new types of qubits) as they weren't interested at all in the engineering aspects of building a quantum computer. Personally I always enjoyed applied research so I think I would have had fun working on "engineering" problems as well, on the other hand I get to work on challenging computer science problems now too and in addition I can mostly decide myself what to work on (as long as I turn a profit, though). Maybe I'll return to QC after my current startup, which I think will turn out better than the last one!


An observer can't tell if you are successful or not. But are not Schrödinger's cat questions typical in quantum computing?


I guess :) Just interesting sometimes to wonder where one would be if one would have chosen differently at a given point in life.


I'm studying psychology and team dynamics and this is such a great example of conflict destroying the relationship. Want to share couple of thoughts on how we can avoid sacrificing the relationship.

-I'm 'The responsible'. Yes, as leaders, we are responsible for the results, at the same time there are people above us who is responsible for the results more than us. The responsibility is never on the shoulders of a single person in the company. When higher-ups support anti-leaders (as in this case), they take on some of our responsibility, and the honest discussion about their expectations is required. There could be lots of reasons for them being not confident in me driving the team, and it would be very beneficial to get their feedback before proceeding. Our assumptions and using formal understanding of the roles is only helping to fuel the conflict.

-The 'I know' trap. We really believe in our understanding of the future, it is always linear and always clear. At the same time, future we understand consists of 2 things: all the bad stuff happened to me before and all the good stuff happened to me before and depending on our feelings right now, we are projecting the mix of those 2 into the future. The question here is where is that 3rd category of all the thing that never happened to me? Where are the things I've never experienced? What if I don't know how this will play out? To me, creativity is something not known and something haven't been done before. Allowing ourselves to not know exactly how it will play out, helps us to keep an open mind and learn from the new experiences.

In other words we are never 100% responsible for controlling everything around us and the world is a bit more complicated than we want to believe. Hoping this was useful! Happy Friday!


It was useful, thanks! I've definitely seen this in tech, esp. the latter case, which I've seen make smart people who deserve better miserable.


Reading the transcript, he comes off as a shitty boss that refuses to give people independence in this group to try new things. It's his way or the highway. He hides this behind him being a "rare" kind of optimist when really he seems to just ignore the important social factors in being a leader.


This is not my reading at all.

If you have ever spent time in a FAANG research org - it can be incredibly frustrating to see so much lack of focus. Everybody does what they want. This is not an issue for most e.g. ML research (my org) because there is so much money to go around, and researchers will leave if they are too constrained.

For hardware/critical projects or customer retention, this is very much an issue. Just look at how the TF ecosystem developed API wise.

There can literally be a project with enormous importance for an ecosystem where multiple teams under one VP do the same thing slightly differently. VPs do nothing and have no opinion on the details. Internal and external users get frustrated because there is no clarity.

If you do a project where you have resources for just one shot in a given timespan, this simply won't cut it.


I think there is a way to set a strong objective s and give freedom for the solution. Generally I’ve found that success comes from limiting objectives: one part focus, and one part humility, saying no to anything ill-defined or potentially impossible. Those are the things that the marketing engine is built on. The difference between the audience and the magician is that the magician doesn’t believe in magic. There are many more audience.

Martinis seems like a very focused leader. The problem chosen was very specific to what they knew they could do with the technology, assuming only that the theory works. He had to say no to business units asking for application value, no to easier paths to publication, no to other people trying to build their careers on a different problem, because he needed everybody’s best ideas on the one problem. But many people can’t adapt to a new problem, and many leaders can’t tell the difference between an idea for solving the problem, and an idea for solving a different problem.

The wiring thing sounds petty, but maybe it was really a divergence. I don’t know. It sounds like part of the problem is that he achieved the goal and didn’t know what to lead with next.


The wiring thing highlights the problem I think. He was the hardware manager but he didn't actually have control over the direction of the hardware team. That's a terrible position to be in.

Also, don't underestimate wiring in a quantum computer. I don't know the details of Google's system but Quantum computers uses analog RF pulses to interact with the Qubit system which are then later digitized. The design of a wiring system in a Quantum computer is similar to what you would see designing a high end phased array radar system with the need for everything to be phase matched, highly stable, and temperature controlled except that a quantum computer is also extremely susceptible to cross talk at the interfaces. Wiring a system of Google's scale can easily run into the $100k+ range in just material and manufacturing costs. If the other person didn't have any experimental experience like Martinis says, then he would almost certainly run into problems.


Yes. People usually leave jobs because their boss doesn't meet their needs. Hartmut hired Martinis to be a manager and then didn't give him authority to do his job. Martinis tells how he had several meetings with Hartmut about this problem but nothing was done. It's clear to me that Hartmut failed in his role.


Yes, we built an IC just for the wiring. I think I have a guess at what the argument was about, essentially expert heuristics versus learned optimization. It's been my experience with Google on other RF projects that they prefer naïveté versus expertise. This was explained to me as advantageous for scaling teams and minimizing technical debt, which actually makes sense.


In my experience at honours university level, which may or may not be relevant to the conversation, is that everyone does what they want if the leader figure doesn't understand the topics in depth.

In my case, I am someone who wants to do everything from first principles whereas my study leader at the time unfortunately told me "you can't understand everything". Although that statement is true, you should at least try to understand one thing really as deep as you can. But what bothered me was that if I wanted to understand something deeply, I realised I had to do it myself. So what I can say is that you don't have to be completely unfocused; it could be that your interests and objectives are different in the group.

Long story short, my HonsBSc was a basically a failure and my MSc was a success for these reasons. The funny thing is that I get satisfaction today when I do anything that would have been a great honours project in my topic (signal analysis in bioinformatics).


> Everybody does what they want.

As opposed to what? Everyone having their own personal sort of projects or splitting up in groups of two and three is natural in the academia research groups I’ve been in to.

What is wrong with that?


Because you cannot deliver experimental physics breakthroughs or hardware with teams of two deciding every week what they feel like doing.

Have you ever thought about how much work it takes to deliver hardware? How many teams and processes and planning there are involved from physics level to software stack to deliver something usable? It's insane. And normal processor hardware is something people have experience with. For a physics experiment, virtually everything is custom made and you only have budget for a few fixes or trials on each component. There is just not much room for slack.


> For a physics experiment, virtually everything is custom made and you only have budget for a few fixes or trials on each component.

I'm actually speaking from my experience in experimental physics (solid state) labs. Certainly, you can't just have teams of two entirely deciding what to do independently of the group, but the level of control and dictation Martinis is speaking about seems like a whole different level.

He got pissed just because a theorist tried to come up with an alternate way of doing this wiring and the lab leader didn't shut it down in favor of just him researching the problem.


Smart comment. I work at a large lab that does cutting-edge hardware development. The slow buildup of fabrication and fundamental-physics know-how is hard for someone who has mostly worked on software to appreciate. To realize these breakthroughs, you need to plan at the decade time scale.


TF?


TensorFlow. The API is outrageous.


I'm not quite a dev, but I have always felt there's something awful with the TF ecosystem. I thought I was the only one 'struggling' with it and that probably there was some fundamental insight I was missing.


What is awful with it?


This summed up my experience with it pretty well:

https://news.ycombinator.com/item?id=21216200


Possibly TensorFlow


On the other hand I really liked the interview. He gives straight answers, no bullshit, no need to read between the lines. It's his own version of it but looks quite sincere precisely because he is not painting himself as a hero, just someone who didn't fit with how things are being run.


It's interesting to see your take versus some of the views below like

>The pattern seems familiar, take a succesful, self-propelled creative inventor, typically an engineer kind of person, put that person in a group where they need to submit to ideas of others, and things eventually go wrong

Not saying any view particular is right or wrong since these things usually are tricky


Speaking as a former student of his, he has a good heart but can be quite difficult to work for. It very much is his way or the high way in terms of overall direction. That said, he can be convinced by a sound physical argument. He is above all a great physicist.


Nothing more infuriating for a knowledge worker, than duplicated effort.

No one likes a 'silo' either.

That said, if someone has a proven track record in an area of expertise, it probably makes more sense to organise around them and move overlapping resource on to some another area of responsibility, if only for productivity's sake.

It would only be a bad idea, if no one was allowed to review or challenge the work produced by the aforementioned expert...

Who knows.

We only have one side of the argument here. I'm sure he'll find a better environment for himself very soon.


Having one guy micromanage a large research group is also very inefficient. Nobody (not even Martinis) knows if and how a practical quantum computer can be built.


Welcome to academia. You just have to ignore the bluster.


Unfair to paint academia like that. My interactions with tenured professors in my dep has been nothing but pleasant. They don't have egos (or don't act like it), are more than happy to help, and have been like a family to me my entire academic career. Don't mean to nitpick, and I know there are academics that are huge pricks and depts that are unfriendly places. But it's not true of them all and I'd hate for people to get that view.


You probably aren't going after the same pool of funding dollars. Folks tend to get along well where there is abundance.


This comment is so dissonant with the average experience in academia it reads like sarcasm


I'm not sure where your idea about the "average experience in academia" comes from (I don't even know if such a thing exist, because the very word "experience" implies a subjectivity beyond the scope of surveys or other objective measures), but even given such an average more positive experiences do still exist in the head of the distribution. But concerning the point: even though my own experience in academia hasn't been exactly pleasant, my professors weren't the source of it (the "system" was), and always provided helpful advice and selfless acts of teaching, like the OP described.


Not really. Academia has it's share of assholes, but the probability distribution of encountering assholes is pretty similar outside of academia too.

Thing is, it can be very brutal if you are flaky. It has a very low tolerance for bullshit/lack of expertise and people will not hesitate a second to call others out on it - even in public.


Outside of bio/medical academia, I think the GP is very true to my experiences as well.


> But Hartmut didn’t back me up and wanted to go ahead and try both.

This seems very Google, and not in a good way. I'm reminded of when they tried to launch Allo and Duo on the same day, causing both to flop.

Google is the poster-child of the company that has the resources to try everything, but that lacks the conviction or the leadership to actually make a decision.


Allo and Duo were complimentary apps, their simultaneous release was necessary and had nothing to do with their adaption rate.


Sure, that's the official Google line.

Meanwhile, everybody who actually tried to follow the launch was confused about why there were two, which ones did what, why there was no interoperability, why wasn't hangouts being replaced, "Mobile-first? More like mobile-only," and "Should we even bother? They're going to kill them both off in three years anyway."

It absolutely did hurt the adoption rate -- It would have gone much better if Google management had announced one communication platform and committed to making it their bet going forward.

They were hedging, and it was obvious. Nobody wants to pay switching costs to use a new thing if the creator isn't even confident in their new thing.


> It would have gone much better if Google management had announced one communication platform and committed to making it their bet going forward.

Ironically, Google had this circa 2013 with Hangouts. The mobile app did SMS, im, video chat, voice chat,. you could do all the same on the web or desktop programs, including sending SMS messages from your computer.

For a short time, Google had an iMessage competitor and if they had merely stayed the course, they would have much more usage today then their fragmented explosion of mobile apps today.

Instead, they keep trying to kill Hangouts.


And they've tried once more, by making Meets free for individuals. I loved Hangouts (& still do!), esp the plugins. I don't get their drive towards fragmentation, honestly


I’ve heard that the best way to get promoted at Google is to launch a new product. If true, that might explain it.


I have a theory that plays into that, which is that Google wants new projects all the time, even if they are new projects, because they think it will keep them relevant and prevent them from being seen like the IBM of our day. Hangouts is a perfectly good product, but that's so 2013!


This fable sounds so old and dusty that probably the only people that didn't hear this parrot song are occupying a shout forum like twitter because they never listen.

I dare you to use the HN search function on comments and come up with a day nobody mentioned this internal career ladder inside Google.


Sorry, are you saying it’s incorrect, or that you’re just tired of hearing about it?


Clichés become so for a reason.


Exactly. I think most employees were even confused mid cheers.

Really when we say google today we are 100% not talking about the place we are remembering.


As far as I understood, they had 2 teams making the 2 communication apps. The simultaneous release date would explain that they were competing with each other, and they didn't want the other team to get the head start on the market. It was probably left to the PR department to sort the problem out and make a press release that made the company look sane.

Imagine if Microsoft had an Excel team and a "Spreadsheetr" team...


I for one can imagine a Microsoft that has Skype, Skype for Business, Teams chat, Yammer, Kaizala...


Live Meeting, MSN Messenger, Lync, Office Communicator


> Imagine if Microsoft had an Excel team and a "Spreadsheetr" team...

You mean as in a "Microsoft Office" team and a "Microsoft Works" team?

https://en.m.wikipedia.org/wiki/Microsoft_Works


Meh. They both passed into my brain, confused me for 60s, and were immediately lost from my consciousness. I think Duo might be a failed Zoom competitor, except that it also seems to be a sketch two-factor product. Anyway, I'm waiting for Allo+, whatever that might be.


Were they? Their screwed up the launch if that were the case...


Freudian slip here... I expect the "adaption rate" was indeed as low as the "adoption rate" :-)


did the other guys wiring end up working tho?


Can anyone describe what the research scientist interviews are like in a FAANG company? How are they similar or different than the SWE interviews?


(FAANG researcher):

The first few tech rounds tend to be the same.

Then there is a set of research interviews and usually a research job talk.

The main observation is that hiring researchers, after passing basic tech rounds, is much more about taste and how people subjectively rate someone's research. It's insanely competitive, but in a sense it is also easier than pure SWE because there are not as many whiteboard hurdles to pass. You just can't really leetcode your way into the role, without a good track record with really interesting research most teams will not be interested.


Do interviews tend to include system design, if the applicant doesn't have industry experience?

Been curious about this too: do FAANG companies hire, say, professors with strong research records, but poor coding skills? I always wondered if everyone on these sort of team were all solid coders, or if there were people there that are exceptions.

Based on your experience do FAANG companies tend to be invested in basic research, or more product focused research?


System design is not a typical part of a research interview.

Yes, they do. The rule of thumb is: The higher the level you are hired, the more wiggle room is there. A random PhD grad will need to pass the normal coding screens which are carried out by SWEs who do not know anything about the candidate.

Last q depends on the FAANG org, everything exists. MS research is very pure, Amazon/Apple very applied, everything in between.


Nice that they actually published the transcript.


> The Google plan is roughly to build a million-qubit system in about ten years, with sufficiently low errors to do error correction. Then at that point you will have enough error-corrected logical qubits that you can run useful, powerful algorithms that you now can't solve on a classical supercomputer. And maybe even at a few hundred qubits, with lower errors, it may be possible to do something special purpose.

Can an expert please explain the applications of this? Specifically, are they expecting to break Bitcoin, forge signatures (break SHA), and decrypt old TLS traffic (break Diffie-Hellman) in 5-10 years?


> Yes, that's right. I was a professor at the University of California, Santa Barbara, and we had government funding and were doing quite well. Google got interested and my group and I came over to work at Google, basically because we both wanted to build a useful quantum computer.

So basically US taxpayers were involuntary seed investors who funded the highest initial risk to achieve validation and proceeded to be wiped out at Round A by Google.


Yes that is how it works and it works well. Government should fund pie in the sky research projects and private industry should take that and make products. It is the history of Silicon Valley and it produced quite a bit of wealth don't you think?


Isn't that a good thing for US if it creates a breakthrough and an industry it can dominate? US has a few years of lead and it's best to maximize when it can.

Other large countries' governments are definitely funding their own research groups in this space and while previously in the 80s and 90s trade sanctions were a potent threat, it's become very clear from the experiences of India and China in the 2000s that there's nothing much US can do (short of a MAD nuclear war) to hold any moderately powerful country back.

So it's best to not fritter away the lead and throw what resources it can when there's still time to keep the lead.


> it's become very clear from the experiences of India and China in the 2000s that there's nothing much US can do (short of a MAD nuclear war) to hold any moderately powerful country back.

The US has not tried to hold China back in any way since 2000. More like it did everything possible to help them grow.


Ref: recent trade war and in particular the action against Huawei. Going overboard on Huawei sanctions in particular underscored the strategic and economic threat to the Chinese leadership (who are by and large pretty competent technocrats). So now they're not going to stop at anything to cut the dependency.

My guess is we'll see a competitive server and/or mobile chip out of China Fabs within 3 years, notwithstanding US pressure on ASML and Netherlands to not export EUV equipment.


Those are recent and not really all that severe actions. Compared to the policies of the last 30 years that basically made china what it is.

Even agains huawei they US stopped well short of any serious actions.


No, this is exactly why Universities exist. As idea machines that occasionally turn out something that changes the world. As well as educating our (elite) children.


And that's great, but nowhere does it say that this something should end up as the property of a corporation who will then charge the public for its use.


Can't enslave professors, so its part of the deal. Been happening about 1000 years. And I'm not sure I care to make Universities into profit centers (any more than they are).


Neither is what being proposed nor is a reasonable reading of the above, if Google hires out an entire group and its associated work then the university and federal research funding agencies should maintain equity in its results, many universities have programs for exactly this kind of thing.

Also afaik John Martinis maintained a paid and tenured position at UCSB during his work at Google.


Not sure how that's different at all. Hiring out === freedom of the individual. Not going to support any rule to keep University employees from going to work wherever they please.

Its the idea that the University would 'own' part of a commercial enterprise that's chilling. Means even more commercialization of Universities than at present.


It's not unusual for universities to own part of a commercial enterprise. Stanford owned a chunk of Google, IIRC. More commonly they derive revenue from licensing intellectual property.


Um hm. And its not unusual for Universities to have become commercial juggernauts run by for-profit administrators with 7-figure salaries, at the expense of faculty and students.


And yet the private sector is not impeded by "freedom of the individual" when it comes to protecting investments with NDAs, non-competes all the way to Levandowski-style criminal prosecution.

Obviously noone should be forced to stay where they don't want to be or prevented to work where they want but in this case it's about what they take and bring with them beyond their innate skills.


On top of what others have said (this is how it's intended to work), if the university saw fit to fund the research up to commercialization, or at least a patent, they would be the ones taking the profit.


You can say that about YC or any other incubator and yet they seem keen on maintaining their equity.


Yes! The thrill of funding something and then paying a corporation who took little risk when the product is finally ready.


It's not even partially ready, but that's not the point - it was more credibly possible when it was acquired by Google then it was when it was initially funded by the public.


What would you prefer? That anyone who receives government research funding be barred from entering the private sector later?


That the public maintains its share in the upside ? For example by having Google put in money back into the NSF budget


USCB probably applied for patents that will bring money back once products are actually on the market.


This is called taxes. Maybe not as high as they should be, but such things exist.


well, them and ucsb undergrads


Great, but it didn't actually tell us much of anything. What would have been so bad about trying his way and also somebody else's way? How does no longer being equipped to try his way make anything better? If the other way won't work, how can Google's project make any progress?


What I find weird is that a technology advances because public money are funding much of the research and innovation.

And there comes a private company and gets whole teams and even patents and projects.

Wouldn't it be fair that tax payers also get something in return of their funding?


> Wouldn't it be fair that tax payers also get something in return of their funding?

What's the purpose of scientific funding: to progress science or to make money?

If the goal is altruistic progression of science, then undoubtedly that's best accomplished through free, nearly-unrestricted publication. That includes both work on nonpatentable fields (fundamental science, or advances too remote from commercialization for a patent timeframe to be viable) and not aggressively pursuing the full suite of IP protections for work that may be patentable.

If the goal is to make money, why should there be public funding in the first place? We don't ordinarily expect the government to run business ventures, and treating science funding like a Bell Labs writ large would be just such a business venture.

On the balance, the situation described in this article appears to be close to an optimal outcome: foundational science happened at a university, but a corporation came in with deep pockets and significantly more funding to work on the commercialization process.


Taxpayer funded research belongs to the taxpayers.


>What's the purpose of scientific funding: to progress science or to make money?

What's the purpose of scientific progress if the people can't share in the benefits?

>but a corporation came in with deep pockets

Deep pockets partially because they domicile in tax-havens, and re-contribute the bare-minimum back to the source of the foundational science, forcing that foundational science funding onto the backs of the people, via trillions in student loans.

How is this "optimal"? Do you think we wouldn't be able to search the web without Google?


If we think of public funding as early stage (extremely risky) investment, then it's more like the lottery that angel investors and vcs play: most investments result in little to no public good but some will make up for bad investments.

I definitely think we should push for fairer and more transparent public investment in science, but saying that all downstream benefits from publicly funded efforts cannot be privatized just seems to tilt incentives in the wrong direction.


we get the technology


Oh yeah behind patents and embedded in corporate secrecy that is almost guaranteed to be irrecoverable once the people that created the technology leave or die. I don’t think cooperations are bad and they should be reimbursed for their expenses, but our current way of handling these things leaves a lot of room for improvement if our goal is to advance the cooperative future of humanity (which it should be).


The point of patents is to incentivize disclosure of knowledge in exchange for limited protection. I personally think that public institutions funding research that leads to patents should share benefits of said patents. I believe most operators of those public funds (ie. Universities) actually benefit from this most of the time. At least in major public European universities


Unfortunately, patents don't really do that in practice. For example, it is typical for engineers to be advised not to read any patents at all. Plus, most patents are written in a style that's pretty much opposite to how you'd write if the goal was dissemination of knowledge.


And the jobs and all that comes with a growing economy.


they probably do, often the university negotiates royalties or equity for spinoffs of research done using university resources.


That’s still private money, no? Well, I guess in this case it was a public uni. It in many cases it is not.


So he read a Thiel book recently and it now pervades his every answer to this interview? And his team had problems with his hyper-focusing? Funny coincidence.

Honestly there's pretty much nothing being said in this interview. People don't get along sometimes.


>So he read a Thiel book recently and it now pervades his every answer to this interview?

Yeah that was kind of annoying. Amazing how even a certifiable genius can fall so completely for such meaningless self-help bullshit


John Martinis speaking at a conference on "Google AI Quantum" last quarter.

https://youtu.be/K_H_mbSH45s


Non paywalled access:

https://outline.com/CnEsTN


I still don't know why he resigned


Let me guess: before anyone asked about his employment status he was both working and not working at Google?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: