Oddly, the website lists "M4 Ultra" which however does not exist... Also, it does not account for Apple Silicon chips to have up to 512GB of memory in some cases, but that might be only a limitation of the gathered data.
I only started charging customers in September. Super-linear growth. I launched annual subscriptions and within less than a week > 15% of customers switched.
There is no real difference between "Ray Tracing" and "Path Tracing", or better, the former is just the operation of intersecting a ray with a scene (and not a rendering technique), the latter is a way to solve the integral to approximate the rendering equation (hence, it could be considered a rendering technique). Sure, you can go back to the terminology used by Kajiya in his earlier works etc etc, but it was only a "academic terminology game" which is worthless today. Today, the former is accelerated by HW since around a decade (I am cunting the PowerVR wizard). The latter is how most of non-realtime rendering renders frames.
You can not have "Path Tracing" in games, not according to what it is. And it also probably does not make sense, because the goal of real-time rendering is not to render the perfect frame at any time, but it is to produce the best reactive, coherent sequence of frames possible in response to simulation and players inputs. This being said, HW ray tracing is still somehow game changing because it shapes a SIMT HW to make it good at inherently divergent computation (eg. traversing a graph of nodes representing a scene): following this direction, many more things will be unlocked in real-time simulation and rendering. But not 6k samples unidirectionally path-traced per pixel in a game.
> If you have some issue with that terminology, by all means raise that issue, but "You can not have" is just factually incorrect here.
It is not incorrect because, at least for now, all those "path tracing" modes do not do compute multiple "paths" (with each being made of multiple rays casted) per pixel but rasterize primary rays and then either fire 1 [in rare occasions, 2] rays for such a pixel, or, more often, read a value from a local special cache called a "reservoir" or from a radiance cache - which is sometimes a neural network. All of this goes even against the defition your first article gives itself of path tracing :D
I don't have problems with many people calling it "path tracing" in the same way I don't have issues with many (more) people calling Chrome "Google" or any browser "the internet", but if one wants to talk about future trends in computing (or is posting on hacker news!) I believe it's better to indicate a browser as a browser, Google as a search engine, and Path Tracing as what it is.
Depressing, but it shows the typical faults of most Canadian projects these days. Massive government spend on a project doomed to fail by economic analysis before it's even online; and no takeaways for the Canadian people to actually get momentum going.
If we wanted to do SMRs right, the goal should be to build one or more SMR production factories, here in Canada, where we manufacture N reactors per month, that fit onto train cars, and can be delivered to qualified, secure sites around the world. Instead, we're paying massive cash out to GE Hitachi, and so the end result will never be "the capability of building and deploying SMRs", it will be "4 unprofitable SMRs in a facility and $4.4 billion a unit if we want more of them to lose money on".
Obviously this is doomed to fail; the units should cost like $100M max so they have positive ROI within a few years. If the unit will never beat solar in $/megawatt for operating and fueling costs, and won't pay for its own construction cost before its lifetime ends, it should never have been constructed; the entire thing is catabolic, all of the work and carbon that goes into it is an utter waste. Everyone involved should just do something else with their lives if we're going to approach it this way.
What's the point? Why do such small-minded people get authority over grand projects?
It’s usually about well connected companies lobbying for free money. It’s the sort of thing that keeps Bell and others afloat and guarantees they never have to get competitive.
I'm still half cheering it because at the very least it's still nuclear progress, and will help ensure we still have nuclear energy workers for another generation here. I worry a lot about what's been done to the Atomic Energy Workers in terms of whittling away at our capability to produce good energy workers with tribal wisdom and the Canadian nuclear culture of safety.
Not with the approach we are showing, but if solar was built like this, it would fail too: remember Solyndra? Treating it as a bespoke construction project instead of as a commodity manufacturing project is the fundamental mistake that continues to result in nuclear costing too much.
Fuck's sake, it's just some hot rocks boiling a kettle, we make it out to sound like it's magic but we had the technology for this ~80 years ago. By now we should have the cost of a standard issue nuclear plant down to way cheaper than anything else. Common layout, protocols, processes, software at all of them... could have been complete in 1989, honestly.
But solar isn't built like nuclear. Solar involves parallel exploration of device designs at very small scale, installed with massive redundancy and resilience. Many billions of PV cells have been manufactured. The real cost decline driver is manufacturing automation. Nuclear, even SMRs, have orders of magnitude coarser granularity.
If you want "hot rocks", it's probably much cheaper to just resistively heat them with cheap solar (you don't even need inverters). This could store energy over many months and, pushed to its cost reduction limits this promises to be the final nail in the coffin for any dreams of a nuclear revival.
>But solar isn't built like nuclear. Solar involves parallel exploration of device designs at very small scale, installed with massive redundancy and resilience. The real cost decline driver is manufacturing automation. Nuclear, even SMRs, have orders of magnitude coarser granularity.
Because the level of permitted development without being crushed by onerous regulatory burdens has been absurdly hamstrung on nuclear. All of the issues you add as "but" cases are things that many different innovations in a fluid market for research could have refined. The same has been done for many complex technologies over the decades, yet for nuclear there's always some excuse like the ones you mention. The comment you replied to is right. We're talking about something that since decades ago could have been improved enormously, and hasn't been thanks to a multitude of stupidities.
The United States Navy trusts extremely compact reactors (designed and working despite the DoD's notoriously lax financial and schedule stringency with defense contractors) to power its absolute most important, costly, defense-crucial war machines, and regularly docks them right inside the country's (and world's) largest urban areas, but somehow there's just no way to make nuclear power for civilian use more compact, cheaper and effective?
The regulatory burden argument doesn't explain why renewables are trouncing new nuclear in China. I view it as a universal excuse nuclear fans trot out to explain away inconvenient realities. They also never explain how the regulatory burden would be reduced in a way that doesn't compromise safety. And regulated safety is the price the nuclear industry pays for liability limits.
> Solar involves parallel exploration of device designs at very small scale, installed with massive redundancy and resilience.
I am imagining a field of shipping-container sized units, each of which is a small modular reactor. Probably with solar panels on top ;) Still a few orders of magnitude different, but the idea here is that each unit is small enough that it can be manufactured, so that nuclear plant bring-ups don't take 30 years. Most of the cost is because of the tremendous generational effort involved in just a single project; what does it take to reduce the cost of the plants themselves to the point where they can really shine, economically?
The goal is to have reliable base load power generation so that we don't have to deal with the massive complexity and carbon footprint of battery plants all over the place to deal with peaky generation technologies like solar. I don't believe that that is a solved problem: using tremendous amounts of rare earth materials for limited-lifespan installations that don't even produce energy is possibly not the best use of our resources, considering it's almost all fossil fuel going into those logistics operations anyway, right? EROEI for a battery plant is going to be hard to achieve.
NPPs that small are a nonstarter, due to loss of economies of scale. Even SMRs are creeping up in size now to try to recapture the economies of traditional gigawatt power plants.
There are two different economies of scale at play here; but if that's the case, I can accept things that are bigger than a train container. Even if we decide that they still have to be just as big as existing reactors, the key thing here is that nobody should start the new wave of nuclear by building a nuclear plant. You have to start it by building factories that produce nuclear plants themselves as a commodity item.
Haven't we learned that already in devops? A nuclear reactor today is a big hand-bombed snowflake. I want nukes-as-code.
Your shipping container mention reminded me of The Box, a book that explains how shipping was so erratic, risky, slow, unreliable and incredibly expensive before the standardization into containers. Containers literally changed the world economy.
I think you are onto something. But this requires upfront investment, which alas, politicians are not for.
Thermal storage has very poor discharge rates unfortunately (usually slower than a day), as well as surprisingly high cost once you factor in inefficiencies and turbine cost
As was repeatedly explained in that other thread, thermal storage of the kind described there is inherently a long term storage technology, and this drives the design to minimize capex, not maximize round trip efficiency. The focus on efficiency is fundamentally misplaced there, as it becomes orders of magnitude less important compared to diurnal storage (which batteries appear to be well on their way to dominating.)
Long term storage and diurnal storage are complementary technologies, sort of like the different levels of cache and main memory in a computer memory hierarchy. Combining them appropriately reduces cost vs. using just one of them.
Anyway, the technology as described would produce heat at 600 C for as little as $3/GJ, which nuclear would have a hard time competing with.
You misplaced a decimal point. A MWH is 3.6 GJ, so it's $10.8/MWH.
$3/GJ is about the current Henry Hub price for natural gas, and as you should know cheap natural gas like this is what killed the "nuclear renaissance" in the US.
Sure. 600 C is about the temperature of steam in a coal fired power plant, so one of the use cases here is to take an old coal plant and replace the heat source. It's much higher temperature than the steam in a LWR, so the turbine can be smaller and cheaper. Also, no steam generator is needed as in a PWR.
Yes? That doesn't mean the capex of a steam turbine for this application would be unaffordable, or that this wouldn't have superior economics to nuclear (which also has a steam turbine, and a more expensive one).
This is concern trolling. The key to nuclear economics is speed of construction, and controlling costs, and not caving to safety pearl-clutchers (that is, adding cost and delays for 'safety measures; meant to appease the public, not things deemed necessary by experts and regulators).
But the key is speed. If you tie up $20B for 20 years uselessly, there's no way you can make a profit on anything.
You're just trying to smear a conclusion you don't like with fatuous insults.
The argument that this time, for sure, nuclear will be much cheaper has worn quite thin. Why do you think anyone in power is going to listen that song again?
BTW, do you think the dominance of renewables over new nuclear construction in China is due to "pearl clutching" there?
There's no conclusion or root cause in the article. It just suggests that since the Canadians had cost overruns and delays, it's impossible to build reactors on time and budget.
Yet China has managed to build those plants exactly around those costs and budgets - I have seen this argument so many time, around high speed rail, where Americans failed at infrastructure and deemed it 'uneconomical', then when China succeeded they smeared them for building probably in 'an evil way' or what.
Let me turn your question back at you - if China is doing so well on renewables, why is it that they're still building tens of gigawatts of nuclear capacity with hundreds more planned?
Even the linked article admits that 'Solar GWh' is not comparable to nuclear GWh because if you add the wattage of panels together you get a meaninglessly big number.
If you are planning around a 24/7 available power source, you need to overbuild solar by 20x I estimate, and the article admits, their calculations do not take storage into account (which you simply would not need if you had an always available power source.
China leads on solar panels, equipment and batteries, yet they are the biggest investors into nuclear today, I think that says enough about solar (and wind) not being able to economically substitute for nuclear.
Also a different angle - economics. If you take 20 years to build a reactor, then the interest that investment assuming an 5% YoY, would be ~2.7x the original purchase price. Your yearly profits wont be enough to pay the interest at that point.
You are right - by these standards it makes no economic sense to build a nuclear reactor, but the standards only exist because of the positively lethargic Western work moral.
You seem to misunderstand me - I'm not some nuclear fanboy, but I'm looking for a powerplant solution that's 3 things: universal (unlike hydro), always available (unlike renewables) and sustainable (unlike gas and coal).
It seems that with SMRs, nuclear is finally getting to that state. I would like to ask you - what is your problem with it?
For me I wouldn't like to live next to a nuclear power plant, but I'd overwhelmingly prefer living there compared to a chemical plant - and there are a lot more of those everywhere.
Plants are huge investment of time and effort and I believe the costs mainly come down sabotage to pearl clutchers like the Greenpeace folks who think every plant is going to turn into Chernobyl, bureaucrats with their own loyalties and agendas of preserving a lucrative status quo and a huge civilizational laziness in the West results in a lack of will to get together and see things through in a timely manner.
Living in the UK but being from another EU country, I definitely see that happening. However, a lot of times it is just due to habits, wrongly-placed mistrust, or not being well settled-in yet because, at the end of the day, there are eg. better GPs and worse GPs everywhere in the world, but if you are still "new" to the country you simply do not know which ones are which, so you prefer to go to the ones you know already.
I'm not entirely sure if the UK has a public-private health system. What I do know is that companies offer private health insurance, even though everyone has access to the NHS. That suggests there's a private system in place, one that probably attracts the most experienced and competent doctors and GPs?
About 10% of people have private health insurance, but note that a large proportion of service providers in the private space also work for the NHS.
E.g. my old GP used to provide both private and NHS services (they were precluded by their NHS contracts from providing private services to people registered with them with the NHS).
Many NHS trusts also provide private services, as they are allowed to do so to improve utilisation and supplement their budgets, so in practice this is part of the reason the NHS is so cheap compared to universal systems in similarly rich countries.
Most private hospitals in the UK also e.g. rely on NHS for intensive care, and this, along with relying on the NHS for first-line care (A&E, GP's unless there's a wait, etc.) is also why private health insurance in the UK is unusually cheap, and why private hospitals in the UK are unusually cheap (if you're in the US, and planning elective treatments, it can be cheaper to fly to London and do it here, even factoring in hotels - and some Central London hospitals have hotel suites, and at least one have or had a previously Michelin starred chef because they cater - literally - to high-end international healthcare tourism).
Oh yeah, healthcare in the US is insanely expensive. They have top professionals, but if you don't have money for long, complex treatments, your options are basically: sell your house, or fly to Cuba, Costa Rica, or the UK for treatment.
I always thought linking all the main things not working in the actual world to the alienation caused by too much digital consumption to be wrong/not really making sense. However, gradually, I am getting closer and closer to that conclusion... In your case, what brought you to the stance "Too much social media is what ails the whole world"? What do you think we could do to solve it?
Social Media used to be better when you actually had a connection to the other person. Nowadays it's mostly anonymous or parasocial. All social media sites have drifted to influencer content (TikTok, Meta, Youtube) or to moving the identity of the other person to the background (reddit, HN). The inbetween of early social media with smaller groups of people who know each other has gotten very rare
The other factor is that everyone now knows how powerful social media can be. Remember when we had positive movements like Occupy Wallstreet, the Arab Spring and Anonymous Hacktivism all facilitated by social media? That doesn't happen anymore. Small things like getting traction for a petition still work, but anything that questions existing structures has no chance of succeeding anymore. Instead social media is overrun by bots that simulate broad consensus on many issues
Are you sure this isn't an Eternal September thing where the initial organizers were just an early-adopting minority, now overrun by a majority that actually has broad consensus on many issues? Also, do you actually have any evidence of bot effects? Would you be able to unleash a bunch of bots on Bluesky and make it seem to have a consensus on tariffs being a good policy?
If you go back in history you can find examples of people making the same claims about too much television. Prior to that, too much radio. Prior to that, too much newspaper consumption.
A common thread is that when people complain about too much media consumption, they’re always talking about other people consuming other media. Few people believe their own consumption to be a societal level problem. Almost nobody believes that their sources of media are the bad ones. It’s always about other sources that other people are consuming.
This is why age verification has the most support of these topics: Adults see it as targeted specifically at a group that isn’t them (young people) whose media they dislike the most.
Did you ever consider that all the concerns regarding the negatives of new media might have some truth to them?
Technology is advancing much faster than humans can biologically evolve and very few people seem ready to seriously tinker with the human genome to keep pace.
Perhaps "the feeds" are just the inflection point where the information overload becomes obvious and baseline humans actually need a majority baseline human experience with all of the associated problems in order to prosper?
So, because some people in the past made (to you) incorrect arguments about something, that means anyone in the future making a remotely similar argument automatically has to be wrong? People in 2025 discussing social media have to be "wrong" because some subset of the population supposedly (to you) made a bad argument about radio 100 years ago?
All of that is broadcast / one direction. Social media is two-way. We've never had two-way mass communication. The rate of communication was an order of magnitude different also.
That doesn't make those claims invalid. Too much television is also a problem, and a lot of television content is junk. Tabloid newspapers are a scourge, as are opinion writers whose output often consists of fallacious propaganda designed to maximize confirmation bias.
In what ways? What things would be better without TV and radio? You think they would be more informed? Or harder to manipulate?
People also complained about literacy rates and the printing press, but how would we have been better off without any of these things so far?
Maybe whatever X newest way to communicate is bad, but when the only evidence against it is the same old arguments that failed to hold up to scrutiny over and over again, I see no reason to give it any more prudence than someone claiming carbonated beverages have caused all out problems. There needs to be compelling evidence beyond people complaining about the collective woes of society that have a cacophony of sources and contributing factors.
To me, different and new communication methods only bring a spot light on issues that we already had. Having a town crier instead of a newspaper, radio, or TV isn't going to make me better informed or less likely to have my information manipulated against me. Sure, it limits the number of sources of information, but that doesn't curate the sources of that information any better when I have no control over them.
As the other user said, people have been warning about new forms of media since the invention of writing. It has always been in vogue to be a nay sayer.
But social media is different. For most forms of media, TV, movies, books, radio etc. You had some degree of agency and choice over what you consumed. You couldn't set what a channel or station was playing, but you could change the channel.
You don't choose what you see on social media. You see what an algorithm thinks is most likely to keep you hooked / going.
Our brains only know what's real based on what's in front of it. You can acknowledge something is rage bait, but as you process it, you will still feel some degree of anger / discomfort. You can acknowledge that something is a cherry picked example, designed to tug the sensibilities of users, but it will still tug on your sensitivities.
And so sure enough, as you keep getting rage baited, concern trolled into algorithmic oblivion, it changes your gestalt. Your worldview shifts to one where those are data points, and it starts distorting your perception of reality.
Garbage In. Garbage Out.
Other people have said that it's like electricity consumption. No. This is very much like tobacco. I don't use social media. Even though I get paid to post to it.
However the uncomfortable truth is that many people enjoy what they see in social media, just like they enjoyed the manufactured bait of Jerry Springer and Jeremy Kyle on TV.
organize things offline. political stuff, social stuff, hobbies, exercise. The things that people want, that online life isn't providing. I can't see another option other than waiting until tech is so commonplace that the advances don't interest people anymore.
Why would that matter? I understood the point was to speed up matrix multiplication by doing the adds and multiplies in a different order. Shouldn't matter whether the datatype is int, float, complex, whatever.
One can use this techniques to optimize the multiplier inside the FP FMA unit. However this cannot be used to multiply two floating point numbers as FP arithmetic is not associative.
Pedantic, but they are top of the market in neither since Switzerland is not in the EU, and definitely not in the UK.
But it is true that in Europe, Switzerland PhDs (and professors too) make most. Not just ETH/EPFL as well. UZH (Uni Zurich) has salaries of 50K CHF per year for PhD candidates (with increments every year) -- that's almost 60K USD by your fourth year. This is also true for other universities. And while Zürich is expensive, it is not _that_ expensive.
His take was not really "novel" however, John McCarthy said basically the same thing multiple times in the 90s and maybe even 80s? He would say something along the lines of "If we ever get to an algorithm that expresses general intelligence, we will be able to write that in one or two pages of a manual. Such a book will still be rather long and the rest of the pages will be about how we got to that algorithm and why it took us so long".
I am not an expert on geometry processing pipelines, however Mesh Shaders are specced differently from GS, essentially one of the big problems with GS is that it's basically impossible for the HW, even after all the render state is set and a shader is bound and compiled (and "searched"), to understand how much memory and compute the execution will take, which breaks a lot of the assumptions that allow SIMD machines to work well. In fact, the main advertised feature of GS was to create geometry out of nothing (unbounded particle effects), while the main advertised feature of Mesh Shaders is GPU-driven and efficient culling of geometry (see for example the recent mesh shader pipeline talk from Remedy on Alan Wake 2). It is true that Mesh Shaders are designed also for amplification, and that word has been chosen specifically to hint that you will be able to "multiply" your primitives but not generating random sequences out of thin air.
It is also true however that advances in APIs and HW desgins allowed for some parts that were troublesome at the time of GS not to be so troublesome anymore.
