Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Claude Code now reads AGENTS.md if there is no Claude.md(claude.com ↗)
    136comments
  2. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    190comments
  3. Saving another 100TB of RAM(cloudflare.com ↗)
    34comments
  4. Cloudflare Quick Tunnels(cloudflare.com ↗)
    223comments
  5. Xcode 27.1 Beta Release Notes(developer.apple.com ↗)
    59comments
  6. How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip(ieee.org ↗)
    discuss
  7. US troop deaths during Iran war exceed Pentagon count by at least four(reuters.com ↗)
    32comments
  8. Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)(arxiv.org ↗)
    11comments
  9. Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug(ledger.com ↗)
    45comments
  10. Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash(cactuscompute.com ↗)
    72comments
  11. How to Write with an LLM(sockpuppet.org ↗)
    246comments
  12. OpenJev(openjev.com ↗)
    238comments
  13. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    35comments
  14. Cyclomatic Complexity in C#(ndepend.com ↗)
    9comments
  15. The Implications of Linguistic Illegibility for LLM Security(arxiv.org ↗)
    17comments
  16. Two parallel neural ectoderm progenitors contribute to the developing brain(newscientist.com ↗)
    51comments
  17. C++26: Trivial infinite loops are no longer undefined behaviour(sandordargo.com ↗)
    168comments
  18. From Geometry to Algebra and Back Again: 4000 Years of Papers (2023) [video](youtube.com ↗)
    discuss
  19. Warez: The Infrastructure and Aesthetics of Piracy (2021)(archive.org ↗)
    13comments
  20. Minimal Phone 2(minimalcompany.com ↗)
    150comments
  21. A 1542 papal cipher cracked with simulated annealing(simonklee.dk ↗)
    discuss
  22. How SpaceX streamlined the Raptor engine(construction-physics.com ↗)
    25comments
  23. Inside ZCode: Silently uploading your Git history to the cloud(ferstar.org ↗)
    89comments
  24. A search-and-inference database from scratch in pure Zig(antfly.io ↗)
    16comments
  25. I vibed a proof of Conway's conjecture(overreacted.io ↗)
    175comments
  26. Size-Specialized Memory Allocation(go.dev ↗)
    3comments
  27. The US 'Kill Chain' That Destroyed an Iranian School(bloomberg.com ↗)
    1comments
  28. Korea raises data breach fines to 10% of revenue(koreajoongangdaily.com ↗)
    73comments
  29. Cekura (YC F24) Is Hiring(ycombinator.com ↗)
    discuss
  30. North Korean nuclear test sets off years of earthquakes(science.org ↗)
    148comments

White House: Developing a Bold Decadal Vision for Commercial Fusion Energy

73 pointsby 4y agowhitehouse.gov
78 comments
4y agoHN ↗

How about getting the fission program into 2021 first?

4y agoHN ↗

How do you calculate or reason about that?

4y agoHN ↗

It comes from believing that the chance of fusion reaching something that's useful is very small. If it's sufficiently small, then the expected payoff < the cost of the program, and hence the program is not worth pursuing.

Where you and he differ is in your judgment of how likely fusion programs are to succeed.

I personally don't think ITER is worth pursuing. I'm skeptical about other tokamak efforts as well. Small efforts that might yield more attractive reactors (Zap, Helion) may be worth pursuing, not least because such efforts would be cheaper.

4y agoHN ↗

the idea that ignores the fact that tackling hard technological problems actually generates a ton of economic activity and generates solutions to other domains.

magnetic confinement, lasers, material sciences have all advanced because of these projects and had impacts in other markets.

even if the money invested never ends up with an ITER reactor. we've still come out ahead economically.

4y agoHN ↗

Fission is today commercially uncompetitive: fission-generated power costs substantially more than solar + wind + storage. Every day, fission becomes more uncompetitive. Fission will never be commercially competitive.

D-D and D-T fusion would cost even more than fission, for the same power output. Fission is today uncompetitive. Fusion would thus be more uncompetitive. The longer it takes to get fusion power working, the more uncompetitive it becomes. Tokamak fusion will thus, also, never be commercially competitive.

But it will be able to absorb an unlimited amount of money.

4y agoHN ↗

You're making a lot of speculation with no basis. Not a single model or citation across half a dozen comments. Just baseless nonsense of the same quality I can find on the street corner.

4y agoHN ↗

We have plenty of evidence of fusion consuming money in unbounded quantities -- the people spending it say they could spend a very great deal more -- and no solitary erg of griddable energy out, decade after decade.

The designs they publish for the fateful day they do get anything out are overwhelmingly bigger than a fission plant of similar capacity, overwhelmingly more costly to build, and overwhelmingly more expensive to operate. Leaving aside the plasma physics, no solutions are known for basic engineering questions in construction and operation.

Fission has turned out to be a dead end, for cost reasons: renewables are cheaper, and easier and quicker to field, and cost is still plummeting, making fission less competitive with each passing day. So, anything that costs more obviously has no chance.

All of this is a click away, if you care to know. Nobody is obliged to know, or care. But not bothering says all.

4y agoHN ↗

If and when we have a fusion demonstrator (preferably several of them) operating for a million seconds with a total energy gain of 100 or more, we'll be in a position to make detailed LCOE estimates.

In the mean time, there are arguments from analogy/difference. The market for fusion power is electricity production. (Other possible uses such as hydrogen production, district heat, desalination, or carbon sequestration have been hypothesized[1], but adding fusion to any of those needlessly adds technical risk, and is unlikely to be greenlit.)

At a very high level, fusion electric power plants would be like fission power plants in all economic respects except being more complicated and therefore having higher O & M costs, and that they operate with bigger parasitic losses and startup power and energy requirements. Therefore fusion plants would have a worse return on capital employed than fission plants.

We observe (1) that capital costs are nearly all of the cost of fission plants, and the same would be true of fusion, and (2) that fission power plants are uncompetitive[2] and will almost certainly remain so. Therefore fusion, being strictly less profitable than fission, will not be competitive.

1. ARPA-E review of initial markets for fusion energy: https://arpa-e.energy.gov/sites/default/files/2021-01/105PM_...

2. Lazard's LCOE Analysis: https://www.lazard.com/media/451881/lazards-levelized-cost-o...

4y agoHN ↗

Burden of proof is on whoever insists it is possible. I have not seen anything persuasive. If you have, it is up to you to provide it.

Nobody has ever come forward with any such thing, to my knowledge.

4y agoHN ↗

I strongly disagree. I think, in the long run, investing in sustainable energy production is the most effective way to spend money.

4y agoHN ↗

Money wasted on wills-o'-the-wisp is just wasted, whichever account it came out of.

4y agoHN ↗

So, you think the likelihood of success is irrelevant to whether something should be funded? Are you also supporting research into LENR, perpetual motion machines and pyramid power?

4y agoHN ↗

The declaration says:

"...Fusion – the same reaction that powers the sun – has the potential to be a game-changing technology to help us achieve net-zero emissions by 2050, increase energy security, and enhance U.S. technology leadership...'"

ITER [1] one of the leading scientific experiments, not a prototype, not a first commercial reactor, but just a scientific experiment, is currently already delayed. The original plan was to start in 2035.

They will have a first test phase of the soundness of the Tokamak then stop for 3 to 5 years and the experiment starts in 2035. After maybe 10 years there would be procurement and contruction ( 20 years ?) of the next step. So would be maybe 2070 or 2080, when we can see the first commercial prototype. DEMO [2]

The comes the question if they are economically viable.

Sounds like the Whitehouse needs to get another team advising them on Fusion....

[1] https://en.wikipedia.org/wiki/ITER

[2] https://en.wikipedia.org/wiki/DEMOnstration_Power_Plant

4y agoHN ↗

"Decades of public investment, billions of dollars of new private investment, and major recent scientific advances suggest that now is the time to move boldly to accelerate development of commercial fusion energy."

They are referring to HTS magnets and startups such as CFS and Tokamak Energy. Previous tracks that were railroaded financially have too slow and inefficient of a track for public appeal. CFS' demonstration 20T coil changes the game. That's why this briefing happened.

4y agoHN ↗

Fair enough. But do any of those tracks realistically provide a road-map that make this timetable look like something more than wishful thinking?

4y agoHN ↗

"Fusion – the same reaction that powers the sun – has the potential to be a game-changing technology to help us achieve net-zero emissions by 2050, increase energy security, and enhance U.S. technology leadership."

I agree with everything in this statement except the idea that anything other than inducing a depression will have the US reach net-zero carbon emissions by 2050. It simply will not happen. Fusion has the potential to be an aid in hitting that by 2100.

4y agoHN ↗

CFS is aiming to have an operational demo reactor that is net energy positive in 2025, and a full scale reactor feeding power into the electric grid in the early 2030s.

https://cfs.energy/technology

4y agoHN ↗

It is one thing to aim, another entirely to have a nonzero possibility of hitting. Investment money seems to need only the aim.

4y agoHN ↗

disclaimer: I am not a fusion scientist, nor even a physicist. Just a bystander who has been following this closely for years.

But: Yes. Just one of those avenues is what MIT's Plasma Science and Fusion Center[0] has been talking about publicly since at least 2017 -- here's a fantastic talk[1] by them about how to think about fusion research and engineering, different approaches available, and about what is different in a world of high strength magnetic tapes, which they have subsequently worked to produce at industrial scales in partnership with CFS[2][3], which was founded by people from the PSFC to take advantage of this research. The bottom line is: these new very strong magnets allow way, way smaller tokamaks to hit net positive.

The hard research here is mostly done. The magnets it gave us have cleared a path forward at a far more tractable time and complexity scale than ITER can dream of. The work is now an engineering and industrial process problem, which is well under way: serious scientists and engineers around the world are hard at work at over a dozen different startups, working on reactors that could fire net-positively in well under a decade.

Will those reactors fix our energy problems? Hey, there's still fuel, waste, maintenance (neutrons hit hard!), regulations, etc... -- commercialization and politics are their own whole problems. But I would absolutely take the long bet that we'll have at least one net positive tokamak running before 2030.

0 - https://www.psfc.mit.edu/sparc

1 - https://www.youtube.com/watch?v=L0KuAx1COEk (really, if you're at all interested, this is simply an amazing talk and clarifies a lot)

2 - https://news.mit.edu/2021/MIT-CFS-major-advance-toward-fusio...

3 - https://www.cfs.energy/ -- you can see Bob Mumgaard, CFS'CEO, is listed as a speaker at the white house event

4y agoHN ↗

But: No. Even if every single thing they hope to get working were to work perfectly the first time, there would still be no commercially competitive power out of it.

If they offered the power exactly at cost, they would get no bids. Fission is today not competitive, and nobody involved can promise that Tokamak fusion could ever be competitive even with fission, never mind with what renewable power will be priced at, by then.

So, no, those reactors will not fix our energy, or any other, problem, unless our problem is too much money and not enough things to waste it on.

4y agoHN ↗

You just keep saying this, but never provide a shred of evidence.

4y agoHN ↗

Evidence is all just a click away, if you care to know.

4y agoHN ↗

Supposing you're discussing in good faith, I'll engage with you, and _agree_: for the near term, I do not believe that fusion will be a commercially viable competitor to wind and/or solar. Not within 10, 20, 30, or even 50 years. I also agree that it likely won't beat the dollar-cost of fission power in that timeframe. But I do think it will replace fission for most new power projects somewhere within that timeframe, especially in developing nations.

The expensive thing about fission is not getting a fission reactor to work, nor keeping it running. It's in getting the reactor greenlit at all, bespoke-designed, and handling supply and waste management. In many places, it is simply not an option due to geopolitics. So fission problem is not impractical because it is cost-prohibitive. Instead, it is impractical because it is responsible-waste-management and weapons-proliferation-regulations and spooky-stuff-difficult-political-battle-PR-nightmares-prohibitive. Sadly, very few people a) want a nuclear reactor in town, b) want to shoulder the risk of building, maintaining, and staffing a fission plant for 50 years, or c) want any "risky" nations to have easy access to them. That, far more than dollar-cost-per-megawatt-hour, is why we don't have nuclear plants everywhere today.

Solar and wind are amazing, and obviously essential to any future we have, but today they cannot do everything, everywhere. Further revolutions in power storage and transmission might enable that (like huge flywheels?) but that's not what we're discussing, we're discussing whether fusion a) can work, b) is useful at all, and c) will ever see large scale use for power production. I do grant that it's _possible_ that a flywheel-revolution could outpace a fusion-revolution.

Because of fission's problems, the niche of fusion over the next 30-50 years may be "nuclear power, but safer": it's more expensive, but it can be sold or built anywhere, the fuel control is not so vital/tough, and the waste is handleable on a timescale we can _at least_ wrap our heads around. Wind and solar currently need baseline plants for supplementation, and fusion may be the ticket to turning off the last remaining gas and coal plants out there, or in supplying power to underpowered nations with less favourable conditions for wind and solar. That is wonderful! That is lightning in a goddamned bottle. Literally! Does it have to be the cheapest option, when for many it's the only option?

Past 50 years, well, history would suggest it'll likely get better. Fission would likely be a lot better today (read: cheaper, smaller) if it were safer to begin with, and solar and wind have gotten astoundingly cheap as they've seen investment and use expand over time, so why shouldn't fusion do the same?

So the argument is: we're now likely to get net-positive tokamaks by ~2030. And if we have any net-positive tokamaks by 2030, then they can likely fill a power-niche that fission just can't. And if we spend 50 years making fusion plants and proliferating the technology, by 2080 it might get pretty damn cost effective. I think that's the bet all these smart people are making, and personally, I'm inclined to take it.

4y agoHN ↗

Solar and wind are amazing, and obviously essential to any future we have, but today they cannot do everything, everywhere.

Well, with storage, they can do everything, everywhere. At worst, you make chemical fuels and ship them off to any weird place that has no sun and wind.

And PV is making electricity at $0.013/kWh in Dubai. Fusion will be lucky to come in within an order of magnitude of that.

4y agoHN ↗

The most expensive and difficult thing about fission is, always, getting it built. It is so very expensive that, often, we give up after spending, literally, billions of dollars beyond what had been projected for completion. Other costs are also high, and seem small only by comparison. Fusion projects promise to cost an order of magnitude more than fission to build, and correspondingly more to operate.

Any place where renewables, local storage, and power delivered via transmission line are all unavailable, even if only temporarily, can surely get by on synthetic fuel shipped in from places that have reliable-enough power to make and stockpile that. So, it is hard to understand a claim that extremely cheap renewables "cannot do everything, everywhere." Indeed, we all are long used to relying on liquid fuels shipped in from across the globe, and I do not know why we would abandon the capacity.

I also see no need for extremely expensive "baseline plants" when the cost for storage is falling much faster, even, than the cost of the energy to be stored. Besides the very mature pumped hydro, we may turn the similarly mature liquified-air technology to storage. Factories for iron-air batteries are under construction. Synthetic hydrogen, methane, ammonia and even kerosene are known to be practical. Even long-distance transmission will often be a cheap substitute for drawing down storage stocks.

While chemical storage is less "round-trip" efficient than some other alternatives, it offers the important advantages that, when the tanks are full, the product is always immediately useful industrially, so synthesis equipment need almost never sit idle; and extra tanks are very cheap, anywhere more would be useful. Synthesis efficiency will only ever improve.

So, supposing somebody did manage to get "net-positive" Tokamak fusion as early as 2030, it is hard to imagine who would still care, by then. Do wake me up, though, if you find a way to get aneutronic D-3He fusion going that I can use for outer solar system operations.

4y agoHN ↗

Fusion – the same reaction that powers the sun

This is dishonest, because the fusion reactions that occur in the Sun, and the fusion reactions that would occur in a manmade fusion reactor, are different reactions.

4y agoHN ↗

It isn't. Stellar nucleosynthesis and DT are both fusion reactions.

It's the same type of nuclear reaction. Both stellar nucleosynthesis and DT are more energetic per reactant mass than typical fission reactions and several orders of magnitude more energetic per mass than any chemical reaction.

4y agoHN ↗

They are the same reactions, in the sense than humans and cows are the same animals because they are both mammals.

They are not the same reactions, in the sense they have different reactants and go at much different rates. And for the purpose of this subject, they are different in the sense that "the sun uses fusion" does nothing to imply that "manmade fusion is interesting or practical".

4y agoHN ↗

Obligatory reminder that at least part of the reason fusion is proverbially "always 30 years away" is because we are still funding fusion research below the "fusion never" level, as we have been since the mid-1980s (before, we were funding it at the "fusion never" level) [0].

Based on the estimates given in 1976 on the linked graph, if we were to dedicate roughly $10B/yr to funding fusion—a tiny fraction of, say, the US military budget—it is likely we would have something at least close to a practical application within no more than about 20 years.

[0] https://imgur.com/sjH5r

4y agoHN ↗

That's not a valid conclusion. Those 1976 graphs were based on some assumptions, assumptions that turned out to be wrong. In particular, they assumed tokamaks work better than they do (under those assumptions, TFTR would have reached breakeven). They also ignore very serious generic engineering/economic issues that were pointed out not long after that (they were publicized in the early 1980s by Lidsky and also Pfirsch and Schmitter.)

It's more plausible, I suggest, to reverse the causality here: fusion funding has been restricted because the technology was problematic, not the other way around.

4y agoHN ↗

I hate to flirt with conspiracy on this, but the funding is so pathetic I really wonder if the fossil fuel lobby has something to do with it. Commercial fusion would be the end for fossil fuels (at least within 50 years).

If we do get fusion expect a monster disinformation campaign similar to the one running now against EVs or the ridiculous 5G FUD campaign.

4y agoHN ↗

No, Tokamak fusion just has no path to practical commercial energy production. The program is just a way to throw money to, mostly, military contractors, and to a handful of physicists, including some plasma fluid dynamics physicists. The last might be able to come up with actually useful ideas, eventually.

4y agoHN ↗

This is quite outdated unless you're maybe looking at only NIF or ITER. Significant progress has been made on the ability to generate magnetic fields with higher temperature more practical superconductors, which is one of the key design constraints.

4y agoHN ↗

NIF never had any intention of producing sustainable fission; it was a re-branded weapons program.

ITER and its spinoffs share the failing that, even if 100% successful in all their aims, power produced would be much more expensive than from fission. But fission is already uncompetitive, and gets moreso every day. So, the longer they take to get to their projected endpoint, the farther they get from usefulness. And they are today far from useful. QED.

4y agoHN ↗

power produced would be much more expensive than from fission.

This is always true at first. People laughed at the idea of getting truly significant power from solar and wind only a decade or two ago. First you make something work, then you scale it.

We are reasonably close to this moment with fusion:

https://www.alamy.com/the-first-light-bulbs-ever-lit-by-elec...

4y agoHN ↗

Some things are only true at first. Other things are true forever. The bigger your fusion plant gets, the more uncompetitive it becomes. Your "scaled-up" fusion plant would never deliver so much as one erg of commercially-competitive energy. But it could absorb an unlimited amount of money, in the meantime.

There have been lots of nuke plant projects that successfully absorbed billions of dollars before finally being cancelled. Every cent found a ready pocket. Producing power was not necessary for that. None of the money has ever been given back.

4y agoHN ↗

Pointing to technologies that succeeded is survivor bias. There are plenty of technologies that lost their race, even renewable ones (for example, wave energy, tidal, OTEC, vertical wind turbines, space solar power, ORC geothermal, thermal biomass). And in general, for any given market niche, there can be only one winner. Most technologies lose. It's presumptuous to assume that fusion will be one that succeeds, for example by presuming that learning will reduce its cost, but not the costs of its competitors. It's especially presumptuous when fusion has very significant obstacles, above and beyond just reaching sufficiently high Q.

It would be better for fusion if these obstacles were taken more seriously, and reactors designed to avoid them, rather than just pretending (via sufficiently optimistic assumptions, for example very large reduction in costs on Nth-of-a-kind plants) they will go away.

The Nth-of-a-kind argument didn't work for fission, btw. It's also not clear how it applies to the non-fusion parts of a fusion power plant. Are the turbines going to suddenly jumpstart a new learning curve just because their heat source is a fusion reactor?

Of the commercial reactor efforts I've been looking at, only Helion and perhaps Zap don't immediately strike me as destined to fail (they may still be destined to fail, I just can convince myself of that yet.)

4y agoHN ↗

I finished watching the event yesterday. Here is a list of my takes, in no particular order:

I'm offended that half of the speakers beat the drum of recruitment for a field that has produced more jaded ex-researchers than actual positions many times over. Filling seats is not an issue.

I don't like how they're combining energy security with progressive politics. I'm progressive, but the president isn't and neither are most democrat career politicians. Tying the fate of fusion to something volatile such as public opinion of progressivism provides no benefit to fusion research programs but adds risk. It's done so the administration can simplify their progressive efforts. "Look we gave all you people a summit and 50 million dollars, now fuck off and get us some margin votes". When they're voted out fusion progress may go with it and start being treated as a football. Not needed. Not wanted.

It's annoying that every politician was like "silly physicists don't know how to communicate" as if communication isn't nearly half of a scientist's job. It's like the politicians have never talked to a scientist. I'm sure they have but making jokes about scientists being science automatons with autism makes the politicians seem relatable to a wide range of voter demographics. Smile for the camera.

The DEI advocates viewed fusion under a scrutunous eye without appreciating that it has the potential to be an equalizer. It fucks people over the least in terms of land usage and fuel/waste issues. Not mentioned once. The people equipped to make these points were not asked about them and yet they label it a conversation. A conversation with themselves.

4y agoHN ↗

Supporting institutional fusion research (ITER, MIT tokamak, et al) feels like discretely calculating the sum of the harmonic series: you can keep going for decades, but you'll never get much more than 57.7% (of a viable reactor.)

I'll be excited about the possibilities of usable fusion in my lifetime when SpaceX starts iterating toward a production design.

[0]: https://en.wikipedia.org/wiki/Harmonic_series_(mathematics)

4y agoHN ↗

You think SpaceX is more likely to achieve commercially viable fusion than MIT and ITER?

4y agoHN ↗

Is zero greater than or equal to zero? Yes!

4y agoHN ↗

What an obviously unreasonable position

4y agoHN ↗

Not the person you're replying to, but I read that as "SpaceX pursuing production Fusion reactor designs would be a strong indication of Fusion being viable".

4y agoHN ↗

The corollary, then is "SpaceX NOT pursuing production fusion reactor designs is a strong indication of fusion NOT being viable".

SpaceX is not, in fact, pursuing production fusion reactor designs.

4y agoHN ↗

Subsidize oil while pumping investment into fusion.... sigh.... Talk about using the nation's dwindling resources to pursue news.