>>I think as a a society we (begrudgingly) make allowances for human error. There is zero solace for a death at the hands of an algorithm.
There is zero solace for a death caused by senseless drivers either. With machines and algorithms you can at least hope they would improve in the future. Humans won't. There will always be a small % of terrible ones.
Imo we should allow (and then enforce) self driving even if machines are still worse than humans (they aren't) because that's the only path we have and the sooner we start walking it the better.
>>SpaceX, meanwhile, trades at 16 times that year’s revenue. Take that as the benchmark — they are both companies with wild aspirations and charismatic leaders — and Anthropic could in future be worth $5tn.
SpaceX has a humongous moat in significant part of their business: launch and Starlink. They are so far ahead other players barely register. There are a lot of options connected to that moat. They also are good at building things fast and already have good infrastructure and are in position to build more.
This is not the case at all with Anthropic. Anthropic is very slightly ahead right now and some of their competitors have structural edges that are hard to overcome. For example Google has data and infra, SpaceX has infra an some unique data, Meta has a lot of data, infra and a cash machine to finance it all and operate at loss as long as they please.
That I think is the biggest problem for Anthropic. There is very little moat and competition is already close and very motivated to catch-up.
If I had to bet they will find buyers at 2T. Very liquid prediction markets are at 2.2T right now. I think they face uphill battle to justify that valuation even if AI sector continue to grow fast though.
I'm sure it does, to some extent that is <100%. But the real limit in a city is land. If we want to use it efficiently, we should check what we want to optimize use of that resource for.
IMO, in Berlin the most lacking thing is the supply in affordable housing, so it makes sense to optimize for that directly instead of a side effect of some other optimization.
Why not build something like Alt-Erlaa? [0] [1]
Cities can just build stuff. And if private developers are not as incentivized to invest, that makes it cheaper for the city to do so itself. As somebody from Vienna, it's crazy to me this is even up for debate.
I am not sure how it is in Berlin. I haven't visited in a while. I think it's too general an assessment though.
In many cities the limit is not land but permits - both being permitted to build at all and then build high enough.
>>Cities can just build stuff. And if private developers are not as incentivized to invest, that makes it cheaper for the city to do so itself. As somebody from Vienna, it's crazy to me this is even up for debate.
I hear a lot of good things about housing in Vienna.
Are those city build apartments readily available for anyone? The problem with public housing when I am from is that it turns the city population into lucky few who get access and the rest who not only have to subsidize it but then need to buy on private market themselves.
Building luxury condos help with housing market as well.
I don't think this is controversial (there are tons of studies that support the conclusion).
>>If you let the "market" decide on its own it will always decide on the most profitable items, even if that's not what is needed.
How do you define "not needed"? Businesses will build what they can sell - the very definition of needed.
The problem is that it's difficult to build. A lot of cheaper housing projects don't get build because some restrictions (like height, number of apartments, parking places) that are not necessary.
So I downloaded that report which of course doesn't contain the most relevant information (the questions) but it contains some examples of wrong answers.
I fed the first question to Grok (which they claimed they tested as well) and it answered it correctly in detail.
I repeated it with another one - again correct answer. I then selected the question they said Grok specifically answered incorrectly and it again answered it correctly.
I am sticking with my first intuition: people are terrible at testing tools and probably wanted them to answer incorrectly/not fully (the questions are constructed in a way to make it difficult as well). They also have vested interest in the conclusion (they are financial advisory firm) so there is that to consider.
People reading ft will now think chat boxes are bad at answering financial questions while they are pretty good at it. Zero consequences for spreading fake news for Financial Times there but good for financial advisors I guess.
>>Okay but why do you feel 3 trials say as much as 10,000?
I don't trust them so I've used 3 examples in incorrect questions/answers they have given and I got correct answers. I spend enough time with LLMs to know that if Grok answered it correctly and in detail then it wouldn't be a problem for GPT or Claude either.
The questions are also constructed in a way that it's easy to answer not fully (which they qualify as wrong). LLMs still answer them correctly and in detail though.
They use random numbers so the answers sounds a bit different but in my experience you won't be able to get a wrong answer to a simple question no matter how many times you try it.
>>The problem when you rely on unclear laws that haven’t been modified for 250 years is they haven’t been modified for 250 years.
The Constitution gives you some rights but maybe not this one (to not be searched without warrant at a border). The problem is not The Constitution but the legislative branch that didn't make laws to ban it.
>>what some wealthy farmers from 250 years ago thought, rather than asking what laws make sense today
It's not the job of judges to invent laws or ponder what laws make sense. That would be legislative branch.
> The US has famously failed to update its legal framework via the legislative branch, so It’s upto the judicial branch which seeks to be about interpreting…
Our judicial system is modeled after Judaic precepts (shoftim/shotrim/shomrim) outlined in Deuteronomy 16:18.
There, shoftim refers to judges who interpret and apply the law, requiring deep literacy in the Torah to ensure righteous judgment without favoritism or bribery. Shotrim are the enforcers of the court (sheriffs, etc.).
I don't know where people get the idea that the role of a judge is not to interpret the law. This is literally the only thing SCOTUS does.
It's their job to interpret but not to invent new laws or ponder which laws "make sense".
In particular just because it "makes sense" to extend 4th amendment reading doesn't justify judges just going for it. That would be legislating from the bench.
Every time SCOTUS makes a decision they invent new laws or nullify existing laws. Every case they take is a question about whether certain laws "make sense," and that determination is based entirely on learned, but subjective, personal opinion.
Take for example laws which depended on Roe V. Wade being a Constitutional right no longer existing, and laws created after that right was repealed, because said right didn't make sense to the current Justices (despite being established law for nearly 50 years.) Did the wording of the Constitution change over those intervening years? No. Justices who agreed that Roe was a right died or retired, and were replaced by justices who disagreed, and the Constitution means whatever a majority of Justices say it does. Or the Supreme Court deciding that the 2nd Amendment's definition of "militia" included every American citizen, thus creating a right to personal firearm ownership where none existed before.
Or more accurately, the Constitution means whatever a majority of Justices decide that the hypothetical ghosts of the Founding Fathers, if summoned into a pentagram, would themselves decide. It just happens to be the case that the original intent of the Founding Fathers always correlates to the political bias of the majority. Crazy how that works.
And of course the Supreme Court's power of judicial review itself was never granted by the Constitution. They just made it up because it made sense to them.
Fortunatly there are ways to codify the agreement of the community/nation through the passage of laws and constitutional amendments. Something the US fails to do, as it’s more convenient politically to allow the judges to do it.
I hold BRK stocks and I expect it to underperform the market on average (but also be a very good hedge in case of big downturn).
BRK is in big part short term treasuries then a stock portfolio then a core business (insurance, energy, rail, manufacturing etc.). While it's reasonable to expect the core business to outperform there is no way the whole thing is a favorite to do so because cash is a big drag.
> but also be a very good hedge in case of big downturn
Are you aware that BRK cratered during the 2009 stock market crash? During a crash, except gov't bonds, pretty much everything sells off equally. There is a joke on Wall Street after 2009: During a crash, correlation on all assets goes to one (1.0).
>>So, easier to sell stuff, easier to buy stuff, free to just travel to/from Canada/EU and work there with no extra paperwork, free to buy and sell service from Canada to EU and vice versa.
>>No Euro, no vote in EU, no ECJ, no payments into each others budget.
This version of EU integration is more popular than EU itself in significant part of EU right now.
Well, the Brexit campaign was manipulated by a state actor (or rather: 'a criminal actor disguised as a state'). None of the alt-right nationalism is in the long-term benefits of the country. This is globalism, the end game. You either team up, thereby reducing your costs, or you get swallowed. Ask Ukraine, Tibet, Hong Kong how it worked out for them not belonging to a bigger team. So divide and conquer is a logical move for the opponent, and it happens to align with alt-right. There is also documented evidence of external influence in these. A medium such as Follow The Money has documented such in various of their articles.
> The whole political course of Ukraine in the last ~20 years is mostly about teaming up with EU and NATO, isn't it?
Not, on the NATO front at least, the part between NATO bending to Russian demands not to extend Membership Action Plans to Ukraine and Georgia in 2008 (followed immediately by the Russian invasion of Georgia)—when Ukraine abandoned pursuit of NATO membership—and the Russian invasion of Ukraine in 2014, only after which did it resume pursuit of NATO membership.
The resumption is sometimes tied to the Revolution of Dignity preceding the Russian invasion, and even occasionally claimed to be a reason for the invasion, but the decision to abandon the neutrality position was well after and a response to the invasion rather than the other way around.
Yanukovych was elected in 2010 promising "economic modernisation, greater economic ties with the EU, and military non-alignment." (https://en.wikipedia.org/wiki/Viktor_Yanukovych). His backsliding on those promises is what caused the Revolution of Dignity. I think the OP's statement is roughly accurate.
There is zero solace for a death caused by senseless drivers either. With machines and algorithms you can at least hope they would improve in the future. Humans won't. There will always be a small % of terrible ones.
Imo we should allow (and then enforce) self driving even if machines are still worse than humans (they aren't) because that's the only path we have and the sooner we start walking it the better.
reply