skippyboxedhero

The long-term is considerably easier because in the short-term there are limits on supply so prices aren't really able to move in a way where balance will be found.

In the long-term, the marginal cost of a request is zero so prices will plunge and a lot of the fixed investment will get stuck. The difference between the cost of training a model/infra and serving a request from that model is a clear problem for economics.

What happens with Anthropic specifically is less clear because they, as every company in the same position in the last 150 years has done, are going to push for regulation. A lot of the risk has also been shifted onto neoclouds. The value of the customer relationship is, however, effectively zero. A core part of their strategy has been to build stuff out like CC to differentiate that but, ultimately, that won't work.

The other strategy they have employed is trying to get other companies to stop training new models. Again, seems unlikely to work.

There is no scenario in which they become Google search outside of government intervention. Obviously this is a key part of their strategy now but at best, this will only work in the US and Europe (I would actually say they are much further along in Europe where there is a strong tendency towards corporatism and centralisation of power, they could be "Europe's AI").

This isn't the case. People assume this is the case because only one side is reported for various reasons.

Just to be clear though: the Communist Party of China is not a massive fan of far-right parties, there are certain ideological barriers there.

Russia employed the former left-wing German Chancellor as part of their strategy making an invasion of Ukraine economically viable for them (it is worth pointing out that when right-wing politicians pointed this out, it was received as a crazy conspiracy theory..."liberal voters" distinguish themselves as people who believe themselves particularly unlikely to believe in "misinformation" whilst simultaneously believing in things that make no sense, most people who have a knowledge of politics are required to believe things that aren't true...if they were in touch with reality, they wouldn't vote).

How dare you. It is only the people that I disagree with that are unstable. I am a stable genius. I am furious.

Have you found a way to stop this?

I used that engineering language skill which is good when it uses it...but it doesn't like to use it.

...there are clearly some areas of alignment that they are not looking at.

One problem is managing expectations from non-technical people. You can ship faster but you are also producing tech debt at that same faster rate, and possibly even faster as the gaps in your initial implementation/prompts are filled by Claude deciding to do random shit.

In most orgs, there is also a gap between leadership, even tech leadership and people actually using LLMs every day. And it is challenging for actual devs who are being told that they should be going much faster and the reality which is more mixed. Pushing back on people who are convinced that Claude is God is going to be tough if you are a junior.

I don't feel the job has really changed, I don't write code as much but that time goes into working out the structure...which is what I was doing anyway. But it has changed the general environment, how work is planned, expectations, etc.

You can tell with some of the more recent models that they have been fine-tuned in such a way to produce a lot of descriptions that sound impressive but have no real meaning. In terms of production of semantic text, the recent models are significantly worse than a year ago.

It is not completely useless as a tool but there is still a huge gap in most real-world applications (I work in an org with multiple thousands of devs and I am top 5 AI user consistently, I have used Claude Code since pre-release). If you are building something like an SQL database which has decades of logic tests built by humans then it can work. If you do not have this (i.e. 99.99% of the time) then you still need to be able to issue prompts that give context. And if you start without that context, it will produce shit code in extremely basic scenarios. In particular, humans are far better at producing code that is easy to change (in practice, you need to create unintuitive, non-human rules to govern this i.e. creating many artificial boundaries which would be completely unnecessary for humans).

I am not sure whether the models just capped out in terms of coding or whether they are just trained by people who have never worked on enterprise software but the gap isn't closing. It has just been filled by people who lack the knowledge to identify when they are being bullshitted (the prevalence of consultants suggests that this markets is large).

I have worked on legacy stuff before. In most cases, there is a kernel of reason and logic that can be recovered. I have seen pure AI projects and they are the worst grade of TCS-authored abandonware. They can often do one specific thing but are basically unmaintainable outside of that.

latency lower for satellites, trading firms have looked at this already, no idea if anyone has actually done it

bandwidth would indeed be worse

This article, possibly deliberately because you can read the subtext of exactly what happened with this story, conflates several things knowing that readers won't know the difference.

Transferring assets from Cyprus to BVI? Legal, there is no issue with doing this. What the journalist did, likely upon discovering this was legal, suggested that it was for tax purposes knowing that readers would assume this was illegal in some way...when it is not relevant.

An opinion was then acquired from a KC which, to a normal person, would read as confirmation that something suspect was going on but, if you have experience with this thing, says that in a completely different, unrelated context this would be illegal (it is a bit sad that we live in a world where a journalist will get an opinion from a KC and they will both conspire to produce something that is obviously intended to be misleading to readers).

Backdating asset transfer from Cyprus to BVI? Again, legal. The role of an auditor is not to provide advice on the legality of the transfer. All that happens is that they get some paperwork saying the assets were transferred two years ago, they go...yes, the paperwork is correct. If you think they should be doing anything else, you don't understand what an auditor is supposed to do.

What would be more questionable is if the transfer didn't actually happen two years ago but the auditor would have no way to know if that was the case. This would be easy to determine in other cases but in the case of personal property, there is no way to know.

What likely happened with the story though is that the journalist put in a lot of time, realised at some point that they didn't actually have any evidence (implicating the people they wished to implicate) so wrote the story that Deloitte were to blame. Will be quite interesting to see what happens next because at several points in the article the journalist uses words that do not describe the role of an auditor i.e. Deloitte was "advising on the falsification of records"...when they show an email which clearly shows that Deloitte did not do this (and again, generally, this is not what an auditor is supposed to be doing).

Text generation is not the bottleneck. Does everyone work for Accenture and TCS?

Of course...they want to make money off it so they exploit everyone...but no-one else wants to make money selling it cheaper...

No, the problem is that people see rent-seeking behaviour everywhere where it doesn't exist. If you think you can do it better...do it.

it isn't good enough for an efficient market. this is a common idea amongst people who have never traded. if it was possible to trade this way, it would happen already. this is how some non-equity markets trade. equities do not trade this way because people do not want it. the system we already have has ludicrously low costs.

do not give up before other people

i remember trading a us stock twenty years ago, i don't live in the us, i used a discount broker, the flat commission was something like $20+percentage of trade. this was significantly after there had apparently a massive drop in costs.

there is some kind of amnesia that exists with things that used to cost massive amounts but are now close to free. it is just not possible to explain to someone that in the recent past, most people were just unable to own stocks because of costs.

what do you mean by bad? you want someone to be on the other side of every investment you make...for free? why don't you do this yourself?

Because the decision to leave the company is largely based upon their sudden, newfound financial security. They may give another explanation but the only thing that has actually changed is the identification of bagholders ready to cash them out.

The other issue is that the narrative about safety within these companies is largely a function of the extreme financial incentive.

As an example, Anthropic was an "ai safety" company that has now produced an AI that fails to listen to basic instructions. If you were concerned about safety, would you produce an AI that was unable to follow instructions? You ask a question, it begins executing commands and doing things.

Safety is product to sell to politicians, not consumers.

Not serious.