Hacker News Reader

Comment by pjdemers | Hacker News Reader

The problem with startups building hard tech is it takes a long time to fail. Failures that take decades doesn't work in the startup model. Image working on a fusion reactor. It might take 25 years to fail. The failure may not be that your idea doesn't work, but because some other kind of reactor comes along that is a better business (ie, cheaper). At this point, you are now a world class expert at a technology that is not economically viable. But, you can't transfer that knowledge, or at least most of it. You are also now 50 years old, give or take. You don't have 25 years to work on the next hard problem. You certainly don't have time to fail again. Back in the day when lots of people worked on hard technology, they were paid well to work on the problem; their life time earnings did not depend on creating a viable business. Because of that, they usually earned a pension, too. Because if you failed, you could retire comfortably, or, at least have enough to choose a second career that didn't pay as well. That model doesn't work for startups.

View on HN

Former fusion startup founder here. This is a very good point. My startup took 6 years to fail and much of our competition at the time is still at it 20 years later.

There are a number of things which are also impedance mismatches between the typical SV startup mode and hard startups. One of these is the failure path. When an adtech startup fails, the founders just go get well-paying jobs at a FAANG, or at least that what the perception appears to be. When your fusion startup goes bust, you... get evicted and live in your car for a while. There's no on-ramp for failed founders back into the industry. Heck, if its a hard startup, there may be no industry yet to go back to. So, you have to go find something else to do to pay the bills. The personal cost of failure can be much higher than typical startups. "Our incredible journey" for hard startups is more like, "Well, none of us died".

Another thing which doesn't mix well between the typical SV expectations and hard startups is that all the things which VCs think are advantages to being located in SV are generally disadvantages to people working on hard problems. Doing web software with ads? Lots of people to recruit in the Valley. Doing nuclear fusion... is there anyone doing nuclear fusion in the Valley? How do they afford to live when competing with all that cheap ad money? Perhaps I should locate on Madison Avenue, it has the same "advantages".

Then there is the typical way that VCs filter, starting with step one, "Know someone who can introduce you". That's a lot easier when you are in the same location, working in similar industries. It's a poor filter when most of the people working on a problem live far away, doing stuff very different from everyone you know. There are far fewer serial entrepreneurs when the failure time is measured in decades, so networking is more difficult.

I'm happy to see VCs try to return to working on hard problems. The Valley used to be good at that, before they became a center of the advertising industry. But Valley VCs need to read their history. In the late 1950s/early 1960s, New York was the center of finance and advertising. The idea that somewhere out in the wilds of California would become a center of a new industry with its own finance ecosystem was crazy. But everyone in the early Valley knew what those East Coast people were like. The Valley is now the new East Coast people, the new Madison Avenue. Doing a hard startup in the Valley now is like trying to start Intel in 1960 New York. Hard startups are hard enough, without trying to do that.

How do you maintain momentum at these kinds of time scales?

When I ran engineering at a startup, I tried to structure things such that we deliver something worthwhile at some kind of predictable cadence. It's the best set-up for everyone: the employees feel like they've accomplished someting, the investors feel like there's progress being made. With software this is pretty easy to do (and even so, people often fail to do it).

But with a "hard" startup the gap between concrete deliverables can be years, and while you could put together some intermediate milestones, they'd probably feel contrived. How do you keep people engaged over long timescales like that?

I know nothing about businesses like these, but I assume that the people who need to be engaged with the idea and work know that things come at a slow pace and are fine with it.

It's different than getting someone who wants to see results every day. In that type of work, progress is measured differently.

That's a good question. My first thought was, "We didn't". But that's really just pessimistic memories speaking. My second thought was, "A perpetual state of stress and fear!". Also pessimistic memories, but closer to the truth.

In reality, we had initial seed funding for about a year. We had a number of goals we had established, including vacuum chamber, getting the ion source running, first beam, first neutrons. We fell way behind in the first six months because I made the stupid decision to try to build a duoplasmatron ion source to save money. This cost us about 4 months to make a shitty ion source that was full of leaks and barely usable. Luckily the experienced electrical engineer came up with an RF ion source bought used that did the job well enough. We actually ended up meeting all of our 1st year goals in just over a year if I recall correctly.

By that time the money was running out, so DFJ basically extended the funding by issuing us convertible debt in 3 month increments. Every three months we would have a set of goals mutually agreed upon by the board, which was two of us and Tim Draper. This was pretty stressful for us, as we were living paycheck to paycheck and we had no idea if we didn't meet a goal if Draper would pull the plug. Tim was never anything but supportive but we understood the point of milestones. That is also when things got technically difficult, as everything we had done before was using existing techniques but when we went off-the-map things slowed down a lot.

That went on for about 3 years. Basically the same dynamic as software guys tend to see in the classic death-march sprint. I blame leadership (myself).

Eventually we ran up against some hard limits that wouldn't work with our design. We faced a choice between making an expensive bet on a supersized ion source (an inelegant brute force solution, which I know now would not have worked) or a new risky design (which also would not have worked). Tim was in favor of the brute force design, because he saw that as money being the barrier. I had my doubts about the brute force design, and I didn't want to waste the money on something that would only fix one of the 3 major problems we faced. I tried building the long-shot design. Then a piece of plastic got into the main turbopump and it ate itself. The 2000 market crash happened. Everything fell apart.

Fascinating, and also probably painful to live through. Thanks for sharing that story.

If you attempted another fusion startup today would the odds of success be higher, or is fusion still about as far away as it was 20 years ago?

Fusion used to be aiming at being better than fission, because that was the default post-fossil energy source.

But now renewables are cheaper than that. So the competition has gotten better since then, and continues to improve rapidly, and it's not clear fusion can ever get cheap enough to compete with them.

Another argument is that fusion reactors will require robots to maintain them (even the aneutronic ones will become too radioactive for hands on maintenance). But if you require robots that can do that, why not just develop and sell those robots, which would be generally useful?

People say that about molten salt fission reactors too. The maintenance robots will need maintenance robots. Just sell the robots.

Lots of people still in fission and fusion are assuming trouble with renewables related to intermittency, land use, grid integration, EROI, raw materials, etc. might arise as we approach deep decarb. If that doesn't happen, wonderful! If it does, then it may be nice to have some low-carbon atomic power stations on hand or in work to diversify the risk.

It's basically the nuclear Pascal's wager.

With a big risk like climate change, there's room for investors who choose to take an interest in nuclear.

Fossil fuels still power the vast majority of the world. It's not yet clear that we're out of that woods.

For instance, Japan just announced that they're building 33 new coal plants to replace their atomic power stations. How on earth does one reconcile this with the game-over-renewables-are-already-cheaper-than-everything popular narrative?

https://www.nytimes.com/2020/02/03/climate/japan-coal-fukush...

If the climate change situation was really as dire as we're being told, we'd have nuclear coming out of our ears by now, and nothing whatsoever would be able to stop it. It doesn't have to be ultra-expensive if it's not bespoke every time you build it. The investment into fusion would also be 10x what it is today, and we'd already have a viable version of it, after a Manhattan-project style, do-or-die effort.

What's puzzling to me is why the opposite seems to be happening: some parts of the world shut down completely carbon neutral stuff and build things that are going to emit gigatons of CO2 for decades to come.

If we are to reach point of no return (or worse, have already reached it), the focus should be on building such an abundance of energy generation capacity that it would not only replace all carbon-emitting energy generation, but also help power atmospheric carbon capture at a scale that we'd need to avert the climate disaster.

CO2 has a logarithmic effect on temps. Doubling the current PPM will have half as much of an effect. I don't think the situation is dire or critical at all. It's mild. And people instinctively believe this even if they overtly signal that they don't.

> And people instinctively believe this even if they overtly signal that they don't.

I think it's rather condescending to literally claim you know people's beliefs, even when they vehemently voice otherwise.

Yes, because I look at what people do rather than what they say.

For your next trick, you will claim that what people really did before the Clean Air Act proves they want air pollution.

Or, maybe, individual actions and collective actions are different things.

> If the climate change situation was really as dire as we're being told, we'd have nuclear coming out of our ears by now, and nothing whatsoever would be able to stop it.

The problem is that people aren't necessarily great at investing for the long term, especially if they think they'll be dead by the time the benefits arise.

> What's puzzling to me is why the opposite seems to be happening: some parts of the world shut down completely carbon neutral stuff and build things that are going to emit gigatons of CO2 for decades to come.

Because of flawed optics on nuclear, mostly.

A lot of this tbh is probably because of infighting between renewables and nuclear.

Fusion seems like a bit of a black box to throw money into which isn't appreciated. Throw too little and you don't get there. Throw enough and you don't know when you'll get there.

Fission solves a lot of problems but nobody likes it in their back yeard and some countries have gotten a scare of it (japan, germany,etc) It also takes a very long time to build a new planet and there's many people saying gov should invest more in solar/wind/etc instead whilst we're in a time where a lot of nuclear plants are reaching their planned life time or have had it extended for a long time. Hell I think only the UK is planning a new one which is sad to see.

Most renewables aren't viable for base load power. Nuclear is viable only for base load power. Realistically they are competing for different parts of the power utilization and shouldn't be infighting at all. You want nuclear for the amount of power that's always on the grid. You want renewables to be able to handle spiking. The hard part of renewables now is the unpredictable output from varying winds and levels of sunlight. The cost problem is solved. If someone could come up with a way to make renewables very consistent in power output they'd blow nuclear out of the water. One of the reasons hydroelectric is so effective is the amount of water going over a suitable waterfall has a viable minimum (it might sometimes increase after rains but even in dry periods it has a predictable power level).

Not sure why this was downvoted. It's the truth. The two don't (or at least shouldn't) really compete, they (and especially solar) are almost exactly complementary to each other.

It's not the truth. The grid could be powered with no "base load sources" at all. Intermittent + dispatchable could cover everything.

They are both competing for the "inflexible" part of the power supply. They step on each others toes, and do not work well together. A grid with a lot of renewables will have the price of power driven to near zero too often for nuclear to survive.

>If the climate change situation was really as dire as we're being told, we'd have nuclear coming out of our ears by now, and nothing whatsoever would be able to stop it. It doesn't have to be ultra-expensive if it's not bespoke every time you build it. The investment into fusion would also be 10x what it is today, and we'd already have a viable version of it, after a Manhattan-project style, do-or-die effort.

This assumes that behind the business and political worlds we actually personally participate in, there is some ultra-rational global they with long time preferences and everyone's collective best interests at heart. Why would "they" allow all this new coal construction?

There is no "they". There is only "us". When you see your local congressman or MP or mayor speak and think, "what a dunce", as you well might, you need to remember that there is no hidden "adult in the room" who really runs things. This is it. Nobody will come and protect us from our collective decisions.

You are assuming your lack of agency extends to people much higher than you in society, without evidence and against my experience (leaders in organizations have different views -- about their organization's purpose, for example -- than the people under them, views they only share with other leaders, or free agents, if they even ever do share them).

I'd say there are multiple adults in the room, and they sometimes fight, sometimes form alliances. They often do not care what you think is rational, your tribe's word for "good", because they are not of your tribe. Some have long time preferences but "collective best interests" to them means "civilization still stands somewhere on the planet".

Well, "throwaway92384", I am not assuming any such thing. I personally know a number of people who are or have been MPs in the UK and the Netherlands, senior civil servants as well as many people in leadership roles in the private sector. I don't think this makes me particularly special incidentally, the same thing is true for many people who are actively engaged in policymaking.

My comment is based on extensive professional and social interactions with other people in leadership roles. It is very frequently the case that such people, including government ministers simply don't have as much independent power of action as you think.

The original parent comment said that if things were really so bad, then something would be done with the implication that since not much is being done, things cannot be so bad. Since we are not mass-building PWRs and running five parallel crash fusion programmes, it must follow that someone has examined the evidence and decided that the problem of climate change is not so bad and therefore we are not doing those things.

The problem with this view is the assumption that there is a someone or a group of someones who sit back, review IPCC reports and then have the ability to collectively compel the political leaders of the world to act rationally. If there was such a group then their lack of action would indicate that they had made a rational decision not to act.

My point is that there are no such people. The powerless enjoy the fantasy of conspiracy because it means that all it would take to solve our problems is some kind of board room coup of the Majestic 12 standing committee.

The reality is that everyone in the system has to act within their roles in the system. Executives at oil companies will say that they wouldn't mind a carbon tax, as long as it was equally applied to all companies. That's logical as they would continue to have a level playing field with the other oil producers and it will take years for us to get off oil as a species anyway. Oh but they don't have the power to pass an oil tax, that's for government.

Politicians in smaller countries will say, they would love to pass such a tax but unfortunately it needs to be global or it will just erode their competitive advantage without reducing emissions. Only if large trading blocs like the EU and US united on enforcing such a system could it be done.

American politicians might say that they personally would love to but they'd need a majority in the Senate, the House, and the right person in the White House.

None of these people are being disingenuous, they all have an accurate view of the limits of their own power within the system. My point is that there is no "outside the system" which can update the system to work better for large collective action problems or force a solution. It's cogs in the machine all the way up. Even the power of Donald Trump, as leader of by far the world's most powerful nation has substantial limits to his freedom of action.

But there _are_ "such people". The wealthy elite, to be exact. They can order the government to do as they please. If the situation is dire, and their collective several trillion dollars are going to go up in smoke in their lifetime, you can bet they'd ram whatever programs they like through the government to avert that. Take the progressive icon such as Jeff Bezos. $120 billion fortune, trillion-dollar company employing 800K people. All of this is entirely dependent (presumably) on the world order not collapsing by the end of the century. And it's not like the guy is illiterate or anything - he's into space exploration and building hundred thousand year clocks. The amount he pledges to fight climate change? Less than 10% of his wealth. Comparably wealthy Gates, who can't shut up about climate change, spends relatively little on it in his philanthropic efforts. Barack Obama, who should be supremely informed on the topic, bought a property near the ocean for $12M. I refuse to believe that these three people in particular are uninformed. I also refuse to believe that they wouldn't be able to effect change if they really wanted to. And by "change" I don't mean just speaking from a podium like you see them do all the time. I mean real, actual change: fund programs, push sensible stuff through the government, etc, etc.

The rational conclusion I make from all this: the rumors of our impending demise have been greatly exaggerated, if all people with means and influence do to avert it is read bullshit from a teleprompter every now and then.

While I accept your point that very wealthy people do not appear to be acting rationally if uncontrolled climate change will cause as much damage as the forecasts assume, I have two counterpoints:

1) People do act irrationally, even well informed people.

2) I don't actually agree that "they can order the government to do as they please". Saudi Arabia killed one of Bezos' employees and hacked into his phone. What has he been able to compel the US government (headed by someone who hates him) to do about it? Barack Obama did attempt to change US policy on climate change and was not able to because no single person has the ability to move US policy that way.

Khashoggi wasn't Bezos' employee. Even if he were, he'd be one of almost a million people he employs directly or indirectly.

From reading the report it's debatable Bezos' phone was hacked at all. The report seems to be more to provide plausible deniability to whatever Sanchez's relative that leaked Jeff's dick pics to the press.

I don't see what SA has to do with any of this.

> Lots of people still in fission and fusion are assuming trouble with renewables related to intermittency, land use, grid integration, EROI, raw materials

That's whistling past the graveyard, I think. Land use is demonstrably not a problem. Grid integration has long history of naysayers drawing a line in the sand and renewables blowing past it. EROI is just fine and getting better. Raw materials are also not a problem (lies that solar uses rare earths notwithstanding).

Intermittency remains an issue, but it looks easier to solve than making fission (or, to an even greater extent, fusion) practical. Combinations of storage, transmission and dispatch technologies are likely to get us to 100% renewable at a cost fission will have trouble matching, especially if industrial users of thermal energy (which is highly storable) can be brought on board.

I wouldn't say 100%, best case economically is somewhere closer to 60% unless you build huge hydropower. With that, 100% is feasible.

That said, do you know what's the highest uptake (in percentage power capacity) of renewables not including hydroelectric nor biofuels in your country?

Here in Poland it is laughably low. Germany is not doing much better either. So "we theoretically could" does not mean we will.

No, you can get to 100%. A key is hydrogen (or perhaps ammonia) to handle rare long lulls in production. The round trip efficiency on that is lousy, but the capital cost is low (which is more important for that segment). A simple cycle gas turbine plant is only 5% of the cost of a nuclear plant of the same rated power.

Try playing with this online modeling tool:

https://model.energy/

With the plausible numbers there, new nuclear is optimized out of the mix.

This will be more expensive than fossil fuels with free CO2 release, but presumably CO2 taxes get ramped up to the point that's no longer true.

> Raw materials are also not a problem

Knowing very little about the topic, I recently saw an article talking about how the first wave of wind turbine blades was starting to be decommissioned, and running into difficulties with what to do with them. While they're no longer useful on the tower, since they're designed to withstand incredible forces, it's difficult to destroy, recycle, or otherwise dispose of them, and indicated they're ending up in storage.

Do you happen to know more about this particular phenomenon and what the end game is for that? Granted, this is likely less of a concern than the waste that coal creates, or the long term storage of spent nuclear fuel. I'm just curious about it.

I asked a person from one of the biggest wind manufacturers on earth that on Monday evening at an energy networking event. At the moment they plan to either put then all in landfills or chop then up and burn them.

We all agreed that the people will likely not worry about this nearly as much as they worry about burying vastly less but vastly more hazardous spent nuclear fuel (even though spent nuclear fuel in actual fact has a sound safe geological story).

Of course there are peer reviewed studies that agree and disagree with everything you say here. It's just a question of how how reliable the peer-reviewed studies on each side are at this point. This strengthens my point that it's ok for some investors to look into improving nuclear.

It's also extremely popular to say that 100% renewables are a done deal because it's such a great story and feeling.

Most serious deep-decarb clean-energy publications I've seen suggest energy costs go up by $30-50/MWh at deep decarb levels. At those levels, current nuclear already competes. People looking at LCOE today (where the rock-bottom wind/solar numbers depend completely on the hugely-fossil grid for backup and intermittency but hide the carbon emissions with impunity) tend to completely miss this point, their bias that wind/solar are cheaper already being confirmed.

Yet there goes Japan with the 33 coal plants!

You can calculate how much power is needed based on past data. That part is complicated, but not really unreliable, although of course the demand may look different in 20 years.

And yes, we're going to need extra longer-term storage, but less than you'd think (IIRC up to 10% for a large grid).

Nobody knows the cost of that storage. You can calculate an upper bound by using today's prices, but once these things actually come in demand, new ideas are going to spring up and the prices are probably going to fall, just like it did for PV and wind turbines.

It seems to me that investing in nuclear tech today is a really risky proposition.

Given the uncertainties and low worldwide penetration of low-carbon anything, I think there's a reasonable case to be made for the opposite conclusion: it may be risky to not invest in nuclear.

Obviously this conclusion is far less popular.

Fortunately investing in various clean energy tech is not a zero sum game because of various investor's predilection to invest in one sector or the other but often not both.

There was a list posted of the companies around working on fusion recently:

https://www.fusionenergybase.com/organizations/

And an article about why this is the case from the same person:

https://www.fusionenergybase.com/article/the-number-of-fusio...

Hindsight is always twenty-twenty but would it have been possible to create products along the way that would have made enough money to have a longer or maybe even infinite runway? Would such an approach have flown with the VCs?

I am thinking about something like creating an own ion source first, keep building and selling these and using the money to tackle the next problem. Then tackle the next problem. And so on.

Two examples of different hard engineering problems being approached in different ways:

Compare Blue Origin's and SpaceX' approach to developing reusable rockets. Blue Origin is financed by Mr. Bezos. If Bezos stopped that (certainly not intentionally but a bus factor of 1 is bad anyways), Blue Origin would be left with no sustainable business from what we can tell. I am convinced that in the worst case a defense contractor would happily buy in but for the company as a whole it might be over. SpaceX was developing reusable rockets by creating rockets (hard but not as hard as going directly to reusable rockets) first and selling services (that's even better than selling products) to customers to finance the development. They were even using already paid for rocket launches for their experimental landing attempts. That's a lot cheaper than having to pay multiple for a rocket just to see it blow up on landing (or reentry or sometime in between).

The same could be said about Tesla, Waymo and self-driving. Waymo has to reach full self-driving (FSD) to start making money. If they fail, they company is dead or also needs to be sold. If Tesla fails to reach FSD? That would suck, they would have to pay back the customers who bought the FSD package early, maybe deal with a class-action lawsuit because of this and but they could carry on selling cars nevertheless.

Not to mention Tesla is a loss leader for battery manufacturing and power grid systems sales.

Might be a non-answer, but I think this really depends on how much leverage you have in pulling engineering + product culture levers.

In my experience, you need to first define a small set of north star metrics that lead to the promised land, which could be years, if not a decade away. Ideally, this should not be more than 3 so there's focus in the long term. If you define this well, everybody will understand the significance of it and be aligned to hit those objectives. Any increase, exponential or not, is progress the team should feel proud of. This is 80% of the "selling" for long time scale projects.

Once you've done that, all other initiatives should focus on improving only these metrics. Generally speaking, very aggressive milestones and keeping that momentum going, while not punishing teams for missing targets will move the needle faster than the other way round.

> The Valley is now the new East Coast people, the new Madison Avenue. Doing a hard startup in the Valley now is like trying to start Intel in 1960 New York. Hard startups are hard enough, without trying to do that.

I think places like Detroit are promising for this. Lots of cheap real-estate, a local history of technical skills, and serious problems to solve (urban farming for example).

Also due to the internet age, developing nations are also pretty promising for their lack of regulation, and extremely cheap location. But have other serious problems (like availability of smart trained talent).

Detroit has deep, fundamental problems. the weather, the utilities, the failed neighborhoods. You can cut costs significantly just by moving to San Diego and not have to explain to prospective applicants why living in a windswept frozen wasteland half the year is appealing.

I think you missed the point. The valley used to be (still is actually) a windswept burning desert half the year. Which is part of what makes its 60's self comparable to Detroit.

Sorry, but for all of Valley's problems, the above characterization couldn't be more wrong. You might be thinking of a part of California east of I-5, but Valley as it pertains to the tech hub couldn't be more pleasant from the climate perspective.

You might be right, but as a San Franciscan, who had to stop by the office in Palo Alto today, the sun did feel a bit oppressive =)

I feel like a lot of people don't realize how gloomy and cold SF's microclimate is.

San Diego has deep, fundamental problems. the cost of living, the homelessness, the failed neighborhoods. You can cut costs significantly just by moving to Detroit and not have to explain to prospective applicants why living in a feces-riddled yet expensive nightmare is appealing.

I grew up in the Midwest. There's no version of reality where I want to move back there.

San Diego is one of the best places to live in the US. Cost of living, homelessness, and failed neighborhoods are no worse in San Diego than in Seattle or Austin. Benefits are perfect weather, proximity to Silicon Valley VCs, recruiting from the University of California system, and a willingness from talent to relocate to join your startup.

Everywhere has challenges. Attracting real talent in the midwest is challenging. Cost of living in industry hubs is also challenging. There's no perfect location to start a startup.

Source: 7 years at startups in the midwest, the south, Seattle, and now San Diego.

I would be shocked if you found it easier to convince candidates to relocate to Detroit than to San Diego.

lol "feces-riddled" san diego is probably the cleanest city in the country but ok

How is urban farming a serious problem to solve?

Food deserts ARE a serious problem.

Growing [X] amount of produce locally would provide a much-needed source of cheap produce and/or revenue to the local community. There's also general health benefits of the added greenery consuming excess CO2. Even better, captured rain runoff goes back into the soil sooner rather than becoming further concentrated with toxic chemicals before it reenters the ecosystem.

Food deserts are a non-problem arising almost entirely from lack of demand for healthy produce.

> Neighborhood Food Outlets, Diet, and Obesity Among California Adults, 2007 and 2009

> Food outlets within walking distance (≤1.0 mile) were not strongly associated with dietary intake, BMI, or probabilities of a BMI of 25.0 or more or a BMI of 30.0 or more.

https://www.cdc.gov/pcd/issues/2013/12_0123.htm

> The Geography of Poverty and Nutrition: Food Deserts and Food Choices Across the United States

> Using a structural demand model, we find that exposing low-income households to the same availability and prices experienced by high-income households reduces nutritional inequality by only 9%, while the remaining 91% is driven by differences in demand.

https://web.stanford.edu/~diamondr/AllcottDiamondDube_FoodDe...

The most energy and carbon efficient way of providing produce is to grow it where it’s easiest to do so and transport it to where there are consumers who want to buy it.

> There's also general health benefits of the added greenery consuming excess CO2.

Added greenery only consumes excess carbon dioxide if you harvest it and bury it in a very, very deep hole, where neither sun nor oxygen can reach it. Otherwise, it only briefly buffers some amount of carbon.

If it increases total biomass, then it does on average sequester that additional biomass. And even if you burn it, well, the next crop will sequester it again. The main problem with burning fossil fuels is that we dug them up out of the ground, rather than take them from replenishing sources that draw CO2 out of the atmosphere.

It perhaps increases the amount of biomass in circulation - but then again, probably not, as it displaces other food (it's not like the western world is facing food shortages at the moment). It doesn't meaningfully affect anything, given that the problem we have is exponential emissions increase.

The best argument for extra consumable plants would be if they displaced carbon-positive farming. Then, all that carbon added to the circulation would subtract from emissions growth.

That came up recently in connection with a VR-related startup. They've been working for years, and they just canned the project and laid off about 60 people.

It's all in-house code. So now the people laid off have to find new jobs. They don't have any of the buzzwords resume scanners want. Heroku? No. React? No. Unreal Engine? No. Unity? No. SpatialOS? No. WebAsm? No. Kubernetes? No. Docker? No. AWS? No. Javascript? No. Go? No. Rust? No? Android? No. IOS? No. Mobile? No. Tensorflow? No? C++? Yes. Maybe they can get a job fixing legacy C++ applications. Going to a company that has its own in-house technology is a real career risk today.

Meh, I've worked 15 years almost entirely at companies using almost entirely in-house technology. Maybe it's because I don't interview for web-tech roles, but I really have never had any problems getting opportunities, and I don't think I'm unique. I think buzzwords are less relevant if you have any amount of specialist skills or significant experience.

For instance, I constantly head hunted for jobs with Unreal Engine and/or Unity, and the fact that I haven't used them significantly doesn't really affect that much.

I agree with this. It's true that there may be less liquidity in the market outside of web tech stuff, but it's not all jobs. There are plenty of companies looking for people with either a) specialist skills which aren't web or b) good engineer skills generally (agnostic as to specific tech)

Not all companies are playing buzzword bingo.

To counter your experience, mine (in London) was almost exactly what Animats described. I worked on the same application for a decade, we introduced new tech when appropriate, but we also stuck with the fundamentals we chose right at the start simply because they weren't broken. That did mean, however, that when I came to looking for a new role I was _way_ behind other candidates.

Adding my anecdata for London, yeah, if you haven't been tracking the bleeding edge of JS framework and node insanities, people will mark you down because everyone* is looking for "full stack devs".

* Unless you want to do Perl maintenance or C++ for fintech, I suppose.

This defaults to the assumption we are still talking about computing / programming skills and domain knowledge related to that. As another poster mentioned, do you think you would be constantly headhunted if you had a very narrow skillset in something more niche like nuclear fusion?

The lens of the HN community is technology == computer science, when the main tool you have is a hammer. Everything looks like a nail.

> Maybe they can get a job fixing legacy C++ applications.

There's big money in that - those legacy applications are vital to the functioning of big corporations. Specialising in large in-house C++ applications obviously doesn't help develop a career in trendy webdev (and nor should it), but it absolutely gives you some good options.

Hell, there's good money in legacy COBOL applications (both maintaining them, and trying to modernize them.)

Money ain't everything. Folks who are after comfortable pay and a W/L balance don't work for a VR startup.

I used to work on an in-house framework. The trick is to write like "used proprietary JS framework xyz.js, a framework similar to react.js". The scanners will still pick it up.

And of course to keep an eye on the ecosystem. Honestly, if you're a decent developer, learning something you're not familiar (e.g. React) takes a day at best to get at a good enough level to be productive at a new job. It's just a different way to cough up HTML in the end.

Apply to an AR hardware company?

TBH I always thought there should be a calculus for this - value a job for its opportunities, location, stated skill levels required, technologies=marketability etc.

Being attractive to buzzword scanners is a more important career risk. I'd rather be rejected by those automatic systems so I pay more attention to offers from real people that know me from the industry.

Facebook and Google hire tons of C++ programmers. Despite what you may think, the back-end of the internet runs on C++.

Also: If you've done VR development, you have marketable skills. Unreal Engine is C++, and learning the specifics of the engine isn't particularly hard. If you're willing to work under the low salaries and often poor management found in the gaming industry, that shouldn't be too much of a barrier.

There's also gaming-adjacent companies, like Epic, Unity, Roblox, Twitch, and so on.

You're putting a lot of emphasis on ads, but lots of major SV companies have nothing to do with ads... Netflix, Uber, Apple, for example. Maybe you mean "software companies"?

Those are the notable outliers though, most of the B2C Bigtechs are adtech companies: Google, Facebook, Twitter, Snapshat etc. Those adtechs are the cancer of our century.

The B2C business is fortunately more product oriented.

Hmm... I've been in the valley over 20 years now and I never worked at an "adtech" company.

In fact I very much prefer a company where customers pay for the services we deliver - something they want, as opposed to something the user typically does not want: Ads.

Gosh, a lot of what you are talking about sounds like what it is like to finish a PhD that doesn't have a future in academia as well. Sometimes I feel like a failure for choosing software engineering, but I should probably be grateful I have that option.

Many people can only dream of getting a degree that’s an instant ticket to upper middle class life.

this is incredibly true

> Sometimes I feel like a failure for choosing software engineering

Why? what would be better?

Go straight to software engineering, skip the PhD.

I would gladly give 10 years of my life to be in your position. We humans are very weird.

I wonder, was it an honest attempt? After 20 years, you certainly have nothing to hide. I mean, did they expect it to work, or they were just having fun working on a hard problem they loved working on, for someone else's money?

Right. Which is why government research is probably a better model for these world-changing basic technologies than Silicon Valley.

Ugh, can you imagine how bad TCP/IP would have sucked if it had been developed by venture capital funding?

which tcp/ip? There would be 10, all with incomplete feature sets and poor reliability.

"Here at bleh we've developed a revolutionary method for transmitting your TCP/IP packets on-chain using AI to perform network level flow control."

The system would probably store all packets in a block chain and routing would be based on a machine learning system.

"Our router can route an astonishing 2,000 packets per second! No other router can offer you this sort of performance! Blockchain. Go ahead. Ask our competitors if they can match this. You'll find their jaw drops and that they'll tell you that this level of performance shouldn't even be possible on commodity hardware! Our experts hand-tuned the multi-billion parameter neural nets our product uses. The competition wouldn't even dream of putting neural nets of this scale into their routers. That's how far ahead we are on this. You can't get this sort of performance from anyone else in the business!"

     Your Email: (                       )
     [ Yes, contact me and tell me MORE! ]

That made me chuckle. I cringe at exaggerated, buzz-wordy, vague sales-copy as much as the other person, but isn't it shown to work (by Apple et al)? I'd be interested if you or anyone else have examples of to-the-point landing pages / sales-copy with zero exaggeration and gimmicks.

I liked the WhatsApp's back in the day [0], but they had a lot going for them and didn't need any extraordinary claims. MissiveApp [1] is another one that feels comfortable, to me.

[0] https://web.archive.org/web/20090914114259/http://www.whatsa...

[1] https://missiveapp.com

Ahhh, don't tell all our secrets to the world! "It's not ready yet"

and all proprietary

As it is, we've only had half that many versions, all incompatible and multiple in common use today with no end in sight.

If you want a vision of the future, imagine JavaScript running on a cloud - forever.

Or companies which can afford research fellows working on fundamental technologies with a very long view.

Snark aside, that is a good point. Erlang is a nice language and it was developed as a pretty short-term project to fill a specific need. In fact, you might even argue it is a nice language because it was developed as a pretty short-term project to fill a specific need.

> Ugh, can you imagine how bad TCP/IP would have sucked if it had been developed by venture capital funding?

It was. Several times. It was developed by a lot of large companies too. Nobody remember those, for obvious reasons, but people used to use them.

People like to forget. It helps pretend government research and alliances always produce the best outcomes.

Your example may be a bit too old. Even novell netware, token ring etc won't ring a bell to many people.

Try a simpler one instead: Do you guys remember about the various wireless standards before Wifi?

The Orinocos and the Intel 2011 PCMCIA were IIRC among the first ones of the 802.11b - quickly eclipsing everything else.

Also, what about DECT from about the same period? Some companies (mostly in Europe) tried to make an equivalent of Wifi, but using DECT modems. Some devices even made it to the market!

You're talking about the product of standards committees there, not startups. Standards committees are sort of "We're doing this so no one company owns/controls this basic tech, so we can all build new and interoperable products with the basic tech and make money!"

The beauty of standards is that there are so many to chose from! So it's not very different from startups.

Eventually, somethings dominate whether it comes from a standard body or is "branded" as a standard ex-post.

> It helps pretend government research and alliances always produce the best outcomes.

Well, at least on this case government research did produce the best outcome. By far. So far no comparison even makes sense, and people don't even remember there were other attempts.

So you mean, when something fit their narrative (here: government research is better) people forget about other attempts? And counter examples in similar domains (here: wifi) are also dismissed?

Color me impressed!

Wifi is not a counter example. The underlying technology was invented at CSIRO, the Australian government research agency.

Something must be invented somewhere. All I say is government research is not necessarily "better". It just is.

I don't know if it's better or worse, but since governments have a much longer time horizon they can fund more basic research.

Novell had NetWare and IPX. Apple had AppleTalk. There was also Token ring...

Though IPX and friends are children of Xerox’s XNS.

Even the Stanford/Cisco people were thinking "multiprotocol gateway" back then. Much of the push for TCP/IP came from the buy side. I did my networking work when I was at a big aerospace company, and we wanted everything to talk to everything else. Not have DECNET<->SNA<->XNS<->X.25<->TP4 gateways trying to interconvert. The vendors all pushed their proprietary systems.

Even Berkeley didn't get it at first. Early Berkeley TCP/IP only talked to other Berkeley TCP/IP systems. They had to be beaten on to get interoperability to work. 4.3BSD shipped with broken sequence number arithmetic that only talked to itself until we fixed that.

This was back when big buyers had more clout than software vendors. It's hard to do that today.

I truly believe suppliers have more influence on market creation direction than buyers, to all of our detriment.

Wouldn't TCP/IP run on token ring? My understanding that token ring is at the same layer as Ethernet, and thus wouldn't exactly be a competitor for TCP/IP.

Is that true though?

Nobody -- public, private, international -- has had any luck in fusion. I think that has more to do with the problem itself.

But consider "world-changing" technologies like electric cars. The SV model doesn't see so terrible, does it?

(Yes, I know about government grants. It's still a primarily private industry play.)

I saw a controversy recently in an open-source project where someone who packaged up the implementation with a friendly GUI was making thousands a month on Patreon, while none of that was going to the people who actually implemented the hard technical parts.

Credit for these technologies can seem similar - the government does all the basic research (through universities etc), funds the development of the tech through grants and the like, and Silicon Valley just does the customer-facing part and picks up all the credit.

There are a few examples of government agencies which can make many parallel bets, on the understanding that most will fail but a few may pay off big. But, by "a few examples", I actually just mean DARPA. I'm sure there are others, not many though.

Politics does not work well with "most of these projects failed". It's too easy of a target. Anything that failed, is an easy target for a congressman to take aim at as being "wasted", because in some sense that particular investment was wasted, in the same sense that every insurance payment is wasted except for the one right before you need insurance.

But, the logic behind "many long-shot bets in parallel" is inherently more difficult to explain in the political arena, which is why I think VC may actually be better at some of these kinds of things.

> I actually just mean DARPA

NASA is the obvious next example. They pay for a lot of esoteric stuff that once made they don't actually end up needing - but having paid the startup costs - end up contributing to other fields.

I think FEMA also does similar things though at a much less parallel (but a higher volume).

Most of the "regulate and promote an industry" (a dumb idea for structuring them but I digress) agencies also do this. The NRC for example does a lot of research on storing and recycling nuclear waste, but also on new nuclear technologies.

And other military agencies aren't worth leaving out. The army core of engineers pushes a lot of contractors to deliver unique solutions that end up getting resold, see the concrete barriers in Iraq.

IARPA, DTRA, SCO, DIU. The DoD spends like 50B a year in research and development.

NIH and CDMRP grant programs is several additional billion.

NIH is about 40 billion, and does a lot less development.

NASA had, at one time, a "cheaper, smaller, faster" strategy, which they then abandoned the first time it resulted in a mission failure. If they had the 90% failure rate of many VC's, betting on lots of 100x but only 10% of the time startups, they would have been shut down.

> But, the logic behind "many long-shot bets in parallel" is inherently more difficult to explain in the political arena, which is why I think VC may actually be better at some of these kinds of things

This does not mean the VC will actually be better at the task (producing better basic research results), but that it is more realistically going to get the funding in the current climate.

If you look at what's worked for computing in the past -- where the real qualitative changes came from, in terms of basic research -- you have ARPA, PARC, and Bell Labs. They all had the following in common:

- Long-term time horizons of 5-10 years minimum; - Lots of funding with few strings attached - Funded people, not projects - Operated in a given context, but goals were vague (enough freedom to play)

Any funding model that is sensitive to the pressures of quarterly earnings reports is not going to be able to do any of this. Which is precisely why most of the improvements in computing since the early 80s have come as side effects from quantitative improvements like speed, memory, cost, etc.

The current computing culture, and especially the funding model behind it, will not give us the big new inventions we hope for.

>But, the logic behind "many long-shot bets in parallel" is inherently more difficult to explain in the political arena, which is why I think VC may actually be better at some of these kinds of things.

VC backed rockets have already taken longer to take a human to orbit than government money took to take a human to the moon. This is on top of the fact that they have had 60 years of technological improvement.

Simply put VC and capitalism in general is great for giving you turds rolled in glitter, less for building the sewage system to flush the turds out.

The amount of money matters too. I'm sure if we were willing to plough 5% of the US budget into the VC approach we'd see bigger results.

It took 5 years to go from the creation of Nasa to a manned Mercury-Atlas.

We still haven't seen any of the space startups match a manned Mercury-Redsone, which took 3 years.

During the Mercury program Nasa was using on average 0.4% of the federal budget, which was little over 2 billion a year in 2020 dollars.

The big spending was on the Apollo missions which are pipe dreams for any of the current space startups 15 years after their foundation.

You are comparing a different time in the US to now. I dont think this is a fair comparison, everything takes longer now -- buildings, bridges, tunnels. I'm amazed to hear how quickly subways in NYC were built -- these days it takes a decade to build a handful of stops.

I dont think it has to do with startups vs not.

The safety culture and degree of acceptable risk is somewhat different now, from when we were sure we'd get nuked by the Russians in a month. (And also still thought smoking was healthy.)

I’m not sure that’s really true.

The NIH budget is about $40B, 90 percent of which goes to research via about 50,000 extramural grants and the NIH’s own intramural research program. Very little of this is expected to “work” in the sense that it produces an immediate improvement in human health. Same goes for the NSF. The DoE, NASA, and DoD also fund plenty of research beyond the ARPAs (DARPA, IARPA, HSARPA, etc).

The ARPA programs are outliers in that they’re more likely to take larger risks in exchange for bigger rewards, while the NIH (etc) are more incremental. Nevertheless, it’s not true that these programs try to avoid ‘failures’ at all costs.

Well, TCP/IP was the military version: the international government version was the misbegotten OSI protocol stack (yes, before it was in irrelevant model on every IT certification exam, OSI was an actual collection of protocol standards), so governments can produce pretty nasty technologies too.

The importance of government research is waning. Most academics vying for government funding are in various states of despair. The blow up in the size of universities and the number of PhDs etc has created a lot of competition with perverse incentives. In biotech, nothing interesting gets done without very early private investment (pharma mainly, some VCs).

Sadly academia for a while now has been subject to the same short-term funding concerns as businesses but without the same flexibility and options for generating that funding.

I was thinking the exact same thing. Research in academia, research labs, and even in some business environments used to be able to focus on and tackle big problems.

These days, funding agencies/sponsors are incorporating more and more business mentality into the research process and focusing on concepts like "sustainability" etc. with short turn around.

You can't expect short term results on difficult problems. If they were easy, they'd have already been solved.

> At this point, you are now a world class expert at a technology that is not economically viable. But, you can't transfer that knowledge, or at least most of it. You are also now 50 years old, give or take.

> Back in the day when lots of people worked on hard technology, they were paid well to work on the problem

is the model closer to "tenured research scientist" rather than "startup founder or employee"?

It's 1/2 way between. Back in the day all the big "tech" firms had long term R&D. You could spend years working on tech with no commercial viability as long as you could keep making better demos. You were most likely an employee with a PhD and a pension.

But then wall street started demanding quarterly growth and they all cut their R&D departments.

Think Bell Labs or Xerox PARC.

I wonder how different interviews were back then by comparison.

Reputation. Who you knew and who knew you.

And by extension, where you got your PhD and who your advisors were and who they knew.

> Back in the day all the big "tech" firms had long term R&D.

Isn't this true of many large companies today? I don't see Alphabet making any returns on X anytime soon.

Very much agree with most of your comment, but:

> At this point, you are now a world class expert at a technology that is not economically viable. But, you can't transfer that knowledge, or at least most of it.

I strongly suspect that anyone who can become a world class tech expert to the point where they can push at the boundaries of what's possible has transferable skills.

Now, that's a different statement from "can get hired on transferable skills," at least depending on the capacity of hiring insights in the industry at large. But I figure it needs to be said lest we reinforce the problem.

> The problem with startups building hard tech is it takes a long time to fail

No, not really: you know pretty much from the outset if a venture will fail because nobody knows how to do it. Altman has staked his career on an idea that is guaranteed to fail, for example. There is no step-1 towards a definable goal. Not even sure there is a real goal!

There is no example of a "hard startup" like this in the history of the human race. Only silly con valley hubris could imagine a different outcome than total failure. This sort of thing is what governments are for.

There are some problems that just aren't amenable to the startup approach. The Manhattan Project wasn't three guys in a garage living on ramen noodles for a year. Neither was the Apollo program. Some problems just require huge amounts of people, money and time to crack -- and even then, as you correctly note, there's no guarantee you'll actually be the one to crack them. (Ask the Germans who worked on their A-bomb project, or the Russians who worked on their moon shot.)

The problem is that, when you're a VC, the only model you have to apply to problems is the startup model. And when all you have is a hammer, etc.

Nintendo was selling playing cards, Facebook was a student only network, Google wanted to index the whole world, SpaceX sends rockets in space and Uber's founder wanted to get a cab. Hard or soft, there is no rule of thumb.

I would replace the "startup model" with "VC funded startup".

I think the assumption in threads like this that VCs have monopolize/colonized the concept of 'startup' so thoroughly that it's the default.

If we're talking about "other kinds of startups" they mean "non-VC".

This model of making incremental progress still exists, it's called research. Startups couldnt be further from the optimal tools to solve those kinds of problems. In my mind, startups are for problems that can be solved by plugging mature tech A into less mature tech B and turning the crank. If someone's first priority is to make the world a better place and expand human knowledge, they should stay away from the business world.

Reply on news.ycombinator.com