> I always had the impression that the storage requirements for nuclear waste are a lot bigger than what they needed to be
This is debatable and I don't like to -- especially in a public setting -- give an authoritative answer of "yes" or "no" because truth is that neither of these would be correct.
I'd rather state some of the factors involved and give you the idea that this is complicated and that there are very smart people working on these issues and we have good reason to trust them (rather than trying to reason through the problem entirely by ourselves[0]).
So the question about nuclear safety is not a binary one, and truthfully this is true for any question about safety (or most things!). Instead it is about level of acceptable risk. This can be VERY small and to the point where we just treat it as a binary case (but it isn't!) or certain bounds. This type of risk assessment is quite complex, especially with nuclear, and depends of the __type__ of failure. So this also includes a whole other topic of failure design/engineering (e.g. skyscrapers are __designed__ to collapse in on themselves. Not to make them weaker or to cause them to fail, but to control the way that they fail __if__ they fail. You don't want your building falling over and taking out another building (creating a cascading effect), so if it falls in on itself it is less likely to cause more damage. Bridges are another common example if you want to learn more).
Nuclear safety design is quite complex because of all the ways things can go wrong and our great concern for safety (making this a WELL studied topic and why the nuclear industry is one of the safest, and arguably the safest in terms of power generation. Even including the 3 main disasters). Disasters are complex and I'll bet you most of what you've heard about Chernobyl, Fukushima, and 3 Mile are quite limited in accuracy. This isn't because of dishonesty or malintent, but because of the complexity.
So the big questions about nuclear storage is how long we want to have a high confidence in our storage. Certainly we have a large and abnormal margin of safety in nuclear compared to other industries, but the question is if we want this or not (that's not a question that science answers, that's a public policy question and to answer appropriately we need to be aware of our answers to this question for other comparable industries (which means not assuming or "reasoning your way through it." It means spending time)) There are some people who want the high level waste to be safely stored for tens of thousands of years and under strong conditions such as not necessarily understanding English or our conventional warning signs (there's a whole rabbit hole to go down here![1]). But there are others who think it is fine to have a safety solution that is good for several hundred years and that we should use this as we invent better techniques. With this, we actually have the technology and even use it today. This route would allow us to store waste on site and there's discussion of ways to cheaply and effectively decommission plants ontop of themselves. (FWIW, this is the option I'm in favor of) Part of the reason some want this method is because it's extremely reasonable to believe that we actually want this material and as our technology has advanced in the last 70 years, we've learned how to extract much more energy from our fuel (which means much lower waste products, which means lower length of half-lives on material, which also means we can recycle old fuel and turn it into new fuel. To understand this properly we need to understand how the fuel actually works, but this is another long topic. I'll just say that we use a very small portion of it. And I'll also mention that France gets about 17% of its total electricity from recycled nuclear. Not 17% of nuclear energy, 17% of __total__ electricity[2]).
The last thing I want to say is we have to keep in mind how much material we're discussing here. I think people grossly overestimate this. Here's a picture of __all__ of France's high level waste[3]. All 60 years of it. It could fit in a Costco. This video[4] shows Russia's. Again, all 60 years worth. And it is worth noting that Russia isn't doing the same recycling as France, so all that material is technically still fairly valuable. You should compare this to other industries. This is important when we're considering our risk assessment because the footprint of waste storage is still an environmental concern. And if we're producing 240 metric tons of coal waste every year (many train cars worth), it can put this into perspective because all that needs to go somewhere. While numbers aren't quite that high for renewable sources it is non-zero.
And one more thing I need to stress to people. From the perspective of scientists, there is 0 people pushing for an all nuclear power solutions. That's considered idiotic. It is similarly considered idiotic to push an all solar or all wind solution. The consensus is we need a well diversified portfolio of energy generation and that the right choice depends on where the power is being generated and needed. Also in the opinion of scientists the argument is more often about "should nuclear be on the table or not." (I for one think it should be) But being on the table doesn't mean it has to be used. Given the context of this thread, I'm sure many of you can see that there is likely no reason to build nuclear plants in the American southwest where there is bountiful sunshine and wind[5]. The question is far less obvious if we are talking about the Pacific Northwest or New England. The question also matters when 50% of zero carbon energy in the US comes from nuclear and where in some regions it is the __only__ zero carbon producer of electricity[6] (but that's more complicated, because we can build renewables in those locations, but politics is also an issue. But that is why I keep stressing not pretending to be experts in complex topics because that's how people get deceived. And I'll say that both "nuclear bros" and "anti-nuclear-pro-renewable" crowd often get a lot of things wrong, though their hearts are in the right place. Both typically get the proliferation question wrong as well as the thorium and waste questions wrong. I mean both nuclear and climate are quite complicated subjects, so it's probably unsurprising).
So I don't know if this actually helps answer your question or not.
[0] "our" role on the public side is to decide policy, not science. We can litmus test and should definitely be skeptical of claims but at the end of the day there needs to be some trust somewhere and the question is who you're placing it in. The people that publish tons and tons of papers that aren't actually readable to the average person or the person who writes laws and doesn't know a positron from a neutrino. The world is specialized and we need to make sure we know what our individual limits are.
There's going to be an insane rug-pull at the end of this. Nvidia funds coreweave who buys gpus and sells them to Microsoft. Microsoft invested in openai who rents the gpus from Microsoft.
It's really easy to get lost in the technical jargon that the vendors who are selling products throw around, but this article has missed the important part, and spent all the time talking about the relatively unimportant part (data formats).
You need to step back and look from a broader perspective to understand this ___domain.
Talking about arrow/parquet/iceberg is like talking about InnoDB vs MyISAM when you're talking about databases; yes, those are technically storage engines for mysql/mariadb, but no, you probably do not care about them until you need them, and you most certainly do not care about them when you want to understand what a relational DB vs. an no-SQL db are.
They are technical details.
...
So, if you step back, what you need to read about is STAR SCHEMAS. Here are some links (1), (2).
This is what people used to be before data lakes.
So the tldr: you have a big database which contains condensed and annotated versions of your data, which is easy to query, and structured in a way that is suitable for visualization tools such as PowerBI, Tableau, MicroStrategy (ugh, but people do use it), etc. to use.
This means you can generate reports and insights from your data.
Great.
...the problem is that generating this structured data from absolutely massive amounts of unstructured data involves a truly colossal amount of engineering work; and it's never realtime.
That's because the process of turning raw data into a star schema was traditionally done via ETL tools that were slow and terrible. 'Were'. These tools are still slow and terrible.
Basically, the output you get is very valuable, but getting it is very difficult, very expensive and both of those problems scale as the data size scales.
So...
Datalakes.
Datalakes are the solution to this problem; you don't transform the data. You just injest it and store it, basically raw, and on the fly when you need the data for something, you can process it.
The idea was something like a dependency graph; what if, instead of processing all your data every day/hour/whatever, you defined what data you needed, and then when you need it, you rebuild just that part of the database.
Certainly you don't get the nice star schema, but... you can handle a lot of data, and what you need to do process it 'adhoc' is pretty trivial mostly, so you don't need a huge engineering effort to support it; you just need some smart table formats, a lot of storage and on-demand compute.
...Great?
No. Totally rubbish.
Turn out this is a stupid idea, and what you get is a lot of data you can't get any insights from.
So, along come the 'nextgen' batch of BI companies like databricks so they invent this idea of a 'lake house' (3), (4).
What is it? Take a wild guess. I'll give you a hint: having no tables was a stupid idea.
Yes! Correct, they've invented a layer that sits on top of a data lake that presents a 'virtual database' with ACID transactions that you then build a star schema in/on.
Since the underlying implementation is (magic here, etc. etc. technical details) this approach supports output in the form we originally had (structured data suitable for analytics tools), but it has some nice features like streaming, etc. that make it capable of handling very large volumes of data; but it's not a 'real' database, so it does have some limitations which are difficult to resolve (like security and RBAC).
...
Of course, the promise, that you just pour all your data in and 'magic!' you have insights, is still just as much nonsense as it ever was.
If you use any of these tools now, you'll see that they require you to transform your data; usually as some kind of batch process.
If you closed your eyes and said "ETL?", you'd win a cookie.
All a 'lake house' is, is a traditional BI data warehouse built on a different type of database.
Almost without exception, everything else is marketing fluff.
* exception: kafka and streaming is actually fundamentally different for real time aggregated metrics, but its also fabulously difficult to do well, so most people still don't, as far as I'm aware.
...and I'll go out on a limb here and say really, you probably do not care if your implementation uses delta tables or iceberg; that's an implementation detail.
I guarantee that correctly understanding your ___domain data and modelling a form of it suitable for reporting and insights is more important and more valuable than what storage engine you use.
I am currently working with about 100TB data on GCP with BigQuery as a query engine and simple hive partitioning like /key3=000/key2=002/. We are happy because we can run all the queries you want and it is insanely cheap. But latency is reaching quite high levels (it doesn't matter so much for us) but I was wondering, if implementing Iceberg would improve this? Has anyone experience with this?
Overall this kind of architecture is just awesome.
Were these board members brought in under the pretense that they'd benefit by being able to build companies on top of this AI and it would remain more of an R&D center without commercializing directly? Perhaps they were watching DevDay with Sam commercializing in a way that directly competes with their other ventures, and perhaps having even used the data of their other ventures for OpenAI, and on top of it as a board they're losing control to the for-profit entity. One can see the threads of a motivation. That being said, in every scenario I think incompetence is the primary factor.
To your point, no normal, competent board would even think this is enough of an excuse to fire the CEO of a superstar company.
It's hard to believe somehow Ilya went along with it, apparently.
Because everyone else is speculating, I'm gunna join the bandwagon too.
I think this is a conflict between Dustin Moskovitz and Sam Altman.
Dustin Moskovitz was an early employee at FB, and the founder of Asana. He also created (along with plenty of MSFT bigwigs) a non-profit called Open Philanthropy, which was a early proponent of a form of Effective Altruism and also gave OpenAI their $30M grant. He is also one of the early investors in Anthropic.
Most of the OpenAI board members are related to Dustin Moskovitz this way.
- Adam D'Angelo is on the board of Asana and is a good friend to both Moskovitz and Altman
- Helen Toner worked for Dustin Moskovitz at Open Philanthropy and managed their grant to OpenAI. She was also a member of the Centre for the Governance of AI when McCauley was a board member there. Shortly after Toner left, the Centre for the Governance of AI got a $1M grant from Open Philanthropy
- Tasha McCauley represents the Centre for the Governance of AI, which Dustin Moskovitz gave a $1M grant to via Open Philanthropy
Over the past few months, Dustin Moskovitz has also been increasingly warning about AI Safety.
In essense, it looks like a split between Sam Altman and Dustin Moskovitz
There’s a ton of astronomy YouTube channels but this is now my favorite, dethroned pbs space time. The video that converted me was his supernova explanation https://youtu.be/RZkR9zdUv-E - for all the content out there no one else has anything this good about supernovae.
Buy an index fund and when the market is down try and buy more.
If you do wish to do something to actively manage things, try giving Nassim Taleb books a read, or just read about his or Mark Spitznagel's investment strategy. They also keep 97% of their money in an index fund, but the other 3% are slowly wasted away buying far out-of-the-money PUT options on boring stocks that are very cheap to buy because they'll "never happen". And most of the time, they lose that money. But when COVID hits, or airplanes crash into famous buildings, or <insert next surprise here>, those little never-gonna-happen options pay for all the damage to the 97%.
Their core theory is that humanity systematically underestimates the probability of very rare events. So it's not about timing the market, it's about using this exploit in human psychology to reduce or eliminate your "risk of ruin" from very rare events.
I was wondering which generation was the most poisoned by the lead: well it's Gen-X:
"Researchers found that estimated lead-linked deficits were greatest for people born between 1966 and 1970, a population of about 20.8 million people, which experienced an average deficit of 5.9 IQ points per person."
> But normal. The cloud of thoughts gone, I am able to make myself do anything with one thought, I have energy, the aches are gone, the tinnitus is gone, I don't second guess, I don't overthink. I just am and I do. It's amazing.
Have you been evaluated for ADHD? This is one of the classic tells. Methylphenidate is structurally and chemically very similar to cocaine.
Vyvanse is definitely the best when it comes to duration and classical stimulants. Modafinil has even longer duration, but it isn't always as effective at targeting ADHD symptoms. Vyvanse is kind of stupid expensive, but you can get a savings card. It would be $330 for me even with a good Rx plan, and the card knocks it down to $45/month. Sounds like you would also respond well to Concerta (extended release methylphenidate). I don't like it because I apparently have the metabolism of a honeybadger and get like 4-6h of otherwise "all day" meds like concerta. I have to split my vyvanse and take it in two doses to get a full day.
> This is what people like Bezos and Musk must be like
My headcannon is both these guys are ADHD and/or on the spectrum, have a doctor's script for vyvanse, adderall, modafinil, and/or whatever else, and have figured out how to ride the tolerance curve properly to maximize productivity.
The primary feature of the real estate market is that the prices change so slowly that margin calls take years to happen. People are trading houses on 400%-2000% (4x-20x) leverage. Whereas with the stock market and other asset classes, you can't generally get that much leverage, and when you do the risk of getting margin called is perpetual and instant.
So although housing prices don't necessarily increase faster than the S&P500, and that there are many local variables in play, the same 7% increase YoY can really be a 140% increase YoY, while you are also renting out the home for more cashflow.
Liquidity of the housing market has vastly improved over the last 5 years, mostly due to new kinds of lenders and underwriters in the market, with the current year being even more liquid than ever.
The downsides of real estate haven't gone away. Like maintenance and physical presence needed, which is difficult for an individual as the portfolio expands. The physical presence demand - or the need to make it economical for there to be someone else maintaining the property - means that it is difficult to come up with the downpayment for real estate in areas you would actually like to live in, but are fully capable of renting in. So the barriers of entry stay where they are.
This is debatable and I don't like to -- especially in a public setting -- give an authoritative answer of "yes" or "no" because truth is that neither of these would be correct.
I'd rather state some of the factors involved and give you the idea that this is complicated and that there are very smart people working on these issues and we have good reason to trust them (rather than trying to reason through the problem entirely by ourselves[0]).
So the question about nuclear safety is not a binary one, and truthfully this is true for any question about safety (or most things!). Instead it is about level of acceptable risk. This can be VERY small and to the point where we just treat it as a binary case (but it isn't!) or certain bounds. This type of risk assessment is quite complex, especially with nuclear, and depends of the __type__ of failure. So this also includes a whole other topic of failure design/engineering (e.g. skyscrapers are __designed__ to collapse in on themselves. Not to make them weaker or to cause them to fail, but to control the way that they fail __if__ they fail. You don't want your building falling over and taking out another building (creating a cascading effect), so if it falls in on itself it is less likely to cause more damage. Bridges are another common example if you want to learn more).
Nuclear safety design is quite complex because of all the ways things can go wrong and our great concern for safety (making this a WELL studied topic and why the nuclear industry is one of the safest, and arguably the safest in terms of power generation. Even including the 3 main disasters). Disasters are complex and I'll bet you most of what you've heard about Chernobyl, Fukushima, and 3 Mile are quite limited in accuracy. This isn't because of dishonesty or malintent, but because of the complexity.
So the big questions about nuclear storage is how long we want to have a high confidence in our storage. Certainly we have a large and abnormal margin of safety in nuclear compared to other industries, but the question is if we want this or not (that's not a question that science answers, that's a public policy question and to answer appropriately we need to be aware of our answers to this question for other comparable industries (which means not assuming or "reasoning your way through it." It means spending time)) There are some people who want the high level waste to be safely stored for tens of thousands of years and under strong conditions such as not necessarily understanding English or our conventional warning signs (there's a whole rabbit hole to go down here![1]). But there are others who think it is fine to have a safety solution that is good for several hundred years and that we should use this as we invent better techniques. With this, we actually have the technology and even use it today. This route would allow us to store waste on site and there's discussion of ways to cheaply and effectively decommission plants ontop of themselves. (FWIW, this is the option I'm in favor of) Part of the reason some want this method is because it's extremely reasonable to believe that we actually want this material and as our technology has advanced in the last 70 years, we've learned how to extract much more energy from our fuel (which means much lower waste products, which means lower length of half-lives on material, which also means we can recycle old fuel and turn it into new fuel. To understand this properly we need to understand how the fuel actually works, but this is another long topic. I'll just say that we use a very small portion of it. And I'll also mention that France gets about 17% of its total electricity from recycled nuclear. Not 17% of nuclear energy, 17% of __total__ electricity[2]).
The last thing I want to say is we have to keep in mind how much material we're discussing here. I think people grossly overestimate this. Here's a picture of __all__ of France's high level waste[3]. All 60 years of it. It could fit in a Costco. This video[4] shows Russia's. Again, all 60 years worth. And it is worth noting that Russia isn't doing the same recycling as France, so all that material is technically still fairly valuable. You should compare this to other industries. This is important when we're considering our risk assessment because the footprint of waste storage is still an environmental concern. And if we're producing 240 metric tons of coal waste every year (many train cars worth), it can put this into perspective because all that needs to go somewhere. While numbers aren't quite that high for renewable sources it is non-zero.
And one more thing I need to stress to people. From the perspective of scientists, there is 0 people pushing for an all nuclear power solutions. That's considered idiotic. It is similarly considered idiotic to push an all solar or all wind solution. The consensus is we need a well diversified portfolio of energy generation and that the right choice depends on where the power is being generated and needed. Also in the opinion of scientists the argument is more often about "should nuclear be on the table or not." (I for one think it should be) But being on the table doesn't mean it has to be used. Given the context of this thread, I'm sure many of you can see that there is likely no reason to build nuclear plants in the American southwest where there is bountiful sunshine and wind[5]. The question is far less obvious if we are talking about the Pacific Northwest or New England. The question also matters when 50% of zero carbon energy in the US comes from nuclear and where in some regions it is the __only__ zero carbon producer of electricity[6] (but that's more complicated, because we can build renewables in those locations, but politics is also an issue. But that is why I keep stressing not pretending to be experts in complex topics because that's how people get deceived. And I'll say that both "nuclear bros" and "anti-nuclear-pro-renewable" crowd often get a lot of things wrong, though their hearts are in the right place. Both typically get the proliferation question wrong as well as the thorium and waste questions wrong. I mean both nuclear and climate are quite complicated subjects, so it's probably unsurprising).
So I don't know if this actually helps answer your question or not.
[0] "our" role on the public side is to decide policy, not science. We can litmus test and should definitely be skeptical of claims but at the end of the day there needs to be some trust somewhere and the question is who you're placing it in. The people that publish tons and tons of papers that aren't actually readable to the average person or the person who writes laws and doesn't know a positron from a neutrino. The world is specialized and we need to make sure we know what our individual limits are.
[1] https://en.wikipedia.org/wiki/Long-term_nuclear_waste_warnin...
[2] https://world-nuclear.org/information-library/country-profil...
[3] https://x.com/Orano_usa/status/1182662569619795968
[4] https://www.youtube.com/watch?v=_5uN0bZBOic&t=105s
[5.0] https://www.nrel.gov/gis/solar-resource-maps.html
[5.1] https://www.nrel.gov/gis/wind-resource-maps.html
[6] https://app.electricitymaps.com/zone/US-SE-SOCO