arw0n

This is moral relativism at its finest, and just plain wrong. I'm not willing to go so far as to call Anthropic a good player, but they are surprisingly often willing to put their money where their mouth is. Obviously everything can be interpreted as a PR move as well, but we just lack context to know true intentions. Personally I have repeatedly sold being a good org as a PR move, it is the easiest way to do good in a capitalist environment. The success of such a sales pitch significantly relies on the moral values (or lack thereof) of the other decision makers at the company.

With Project Glasswing for example, I'm impressed at how generally well thought out it is, and very much appreciate that they donate a lot of money to OSS. I would have liked them to extend the Project to smaller players as well, but power centralisation is an inherent problem of AI, not something that is unique to Anthropic.

And what is malicious about that ideology? I think EAs tend to like the smell of their farts way too much, but their views on AI safety don't seem so bad. I think their thoughts on hypothetical super intelligence or AGI are too focused on control (alignment) and should also focus on AI welfare, but that's more a point of disagreement that I doubt they'd try to forbid.

Just to be clear, the smartest person is still a minister in Idiocracy, and the whole premise hinges on the idea that the elite still recognizes intelligence as something desirable.

I think they are implying that people with lower intelligence are more likely to be in jail. There's at least three reasons why I would consider this very plausible:

1. Intelligence is associated with higher impulse control.

2. Intelligent criminals are probably less likely to be caught. They might also have a better chance at working to lower their sentences after being caught.

3. It is harder to be successful with a lower IQ (although I would posit it is far from the most important factor), and economic hardship is strongly correlated with crime.

> There are national outlets like ZDF, Tagesschau, Deutschlandradio and regional ones like Norddeutscher Rundfunk and Bayerischer Rundfunk. Each design and present their own programmes.

Well yes, but calling them politically independent is a bit of a stretch. A 2024 study found 52% of board members (Rundfunkrat) have a party membership (~2% of the general population is part of a party). [0]

To take one example you mention, the ZDF-Fernsehrat is dominated by party members (33/60).Notably only by the conservative party (CDU/CSU) and the SocDems (SPD), with 2 green members and 1 member of the SSW. Neither the left party, nor the far right AfD have any representation, despite accounting for roughly 30% of the national vote. Religious communities have signifigantly more representation (9), than the scientific community (0). [1]

Public media was always a tool to help create and maintain a societal overton window of shared truth and identity, and as such very helpful in keeping Germany united and democratic. There was however also always clearly immoral and untrue directions taken for ideological reasons or political convenience, for example the support of Apartheid South Africa til its fall, and the recent biased coverage of Israel. Many other topics as well, like immigration, covid and the war in Russia, are presented in a way that does not align with significant amounts of the german population: We are currently witnessing this overton window breaking apart completely, in other words, German public media has failed in its primary purpose.

[0] https://www.medienpolitik.net/aktuelle-themen/die-politik-is...

[1] https://de.wikipedia.org/wiki/ZDF-Fernsehrat

> "This is the future"

Yeah, that's silly, it is already the present!

Some interesting conversation one can have with coworkers specifically:

1. How should code review and responsibility for code be updated to a) increase velocity, b) keep quality and c) keep reviewers from burnout. There are plenty scenarios in which vibe coding a component in an afternoon is the correct choice, even if it is buggy, insecure, and no one really understands it.

2. Which parts of the codebase work well with code assistants, which don't? Why? What could be changed to make it easier? In my experience, Claude Code sometimes loses its mind on infra topics. It is also not very good at complex, interconnected services (humans aren't either).

3. Which tasks could be offloaded to agents to save everyone time and sanity? - Creatig Jira Tickets from meeting transcripts is an obvious one, collecting and curating bug reports another one.

4. How should we design systems to better work for coding agents? Does it influence our tech choices? Should it influence them?

5. Is AI a net positive or negative for security?

And so much more. The last topic in particular is incredibly important, and things are developing so fast that you can probably have a new conversation on it every two weeks.

See also the new Deepseek paper on engram transformers for some progress in this area: https://arxiv.org/pdf/2601.07372v1

They observe significant gains in factual knowledge retrieval capabilities, but reasoning barely moves the needle.

Mathematics is a language. Everything we can express mathematically, we can also express in natural language. The real interesting, underlying question is: Is there anything worth knowing that cannot be expressed by language? - That's the theoretical boundary of LLM capability.

> People will go out of their way to help you, and to support you. Over time this could very well boost your ego to also become more confident and decisive.

This was my first thought about why the move to online works for men but not women. When I was attractive, people intuitively thought I was competent as well. I could speak with authority on topics not knowing more than the average and people would believe me.

Fo women this might sometimes even be the other way around - more beauty is associated with less intelligence.

Nuclear is significantly more expensive on average. It is in no way an either/or question, you could have both. With 35% of energy being renewables, the country is still very much reliant on global energy markets, and it should indeed be a wake up call. If it was 100% reliant on qatari gas, the situation would be significantly more dire.

From your older comment:

> Our CO2 emissions are like 0.15% of the total and our per-capita emissions are already lower than for ex. Germany. We basically have no industry. If we could wave a magic wand and completely stop our CO2 emissions it wouldn't move the needle on global warming at all.

Well, it would move the needle by 0.15%. Portugal has 0.11% of the global population by the way.

For any command, you can create an 'alias' in your shell config. That way you can get rid of the prefix.

Ruff is nice, but not important, uv is one of the few things making the python ecosystem bearable. Python is a language for monkeys, and if you don't give monkeys good tools, they will forever entangle themselves and you. It is all garbage wrapped in garbage. At least let me deploy it without having to manually detangle all that garbage by version.

I'm done pretending this is a "right tools for the right job" kind of thing, there's wrong people in the right job, and they only know python. If no one self-writes code anymore anyway, at least use a language that isn't a clusterfuck of bad design decisions, and has 1 trillion lines of code in the collective memory of people who don't know what a stack is.

Let's see how it plays out. My current assumption is that degrees and CVs will become more important in the workplace. Things like good architecture, maintainability, coherence, they are all hard to measure. A true 10x developer without a college degree will lose to the PhD without any hard skills. And these types only speak python, so they will instruct the AI to type python. Or maybe they'll vibecode rust and elixir, I don't know. But the cynic in me strongly thinks this will make all our bullshitty jobs way more bullshitty, and impostors will profit the most.

Many professors view teaching as a secondary obligation. Even if they don't it takes more time to learn to teach something than just to learn it. Our field is moving so fast that outside of the major innovations, it would be quite difficult to keep up being a good teacher on everything, while also doing research, and doing the actual teaching. In addition, most new tech isn't very interesting, or useful. Like every couple of months I'm getting another peak at SOTA Python or JS and the "innovation" is just another layer of duct tape that doesn't really improve much.

Cool tech usually also sees faster adoption in academia. Rust courses where offered at the uni I went to back in 2017 for example. According to my friends still involved with uni, there was also a strong shift towards more data science/engineering and HCD since then, both fields that saw major practical improvements.

The trick is to keep a layer of management or engineering below you that can be blamed if things go wrong.

This is assuming a coherent national security strategy, which is unlikely. We know a lot of generals disagree with the attack on Iran, and none of the geopolitical experts I trust think it is a good idea, be they conservative, realist, liberal, leftist or something else.

There's a number of reasons this is happening now that I think are more plausible than American interest:

- Saudis want Iran weak as they are primary geopolitical rivals. There are deep ties between the Saudi dynasty and the Trump dynasty. Without Iranian support, the Houthis will have a much tougher time. (Although they should not be underestimated regardless. They are not an Iranian proxy, but an ally, and field one of the strongest armies in the whole region.)

- Israel wants Iran weak, and pro-zionism is a strong wedge in American politics. Again, there's also a lot of personal business interests involved. Iranian allies and proxies are the chief causes of grief for Israel's expansionist agenda, and a very credible threat to their national security.

- This war conveniently moves the headlines away from a faltering economy, the Epstein files, and ICE overreach. There's probably hope that it will improve chances with the 'war president bonus' in the mid-terms. It could also be a convenient cover for and excuse to increase rigging in the elections.

Expecting positive regime change after bombing a school full of little girls is... naive. This is not how you turn an enemy into a friend.

There seems to be a massive misunderstanding here - I'm not sure on whose side. In my understanding, if the DoD orders an autonomous drone, it would probably write in the ITT that the drone needs to be capable of doing autonomous surveillance. If Lockheed uses Anthropic under the hood, it does not meet those criteria, and cannot reasonably join the bid?

What the declaration of supply chain risk does though is, that nobody at Lockheed can use Anthropic in any way without risking being excluded from any bids by the DoD. This effectively loses Anthropic half or more of the businesses in the US.

And maybe to take a step back: Who in their right minds wants to have the military have the capabilities to do mass surveillance of their own citizens?

Google wasn't bleeding money like crazy at the time. Google was operating in a post-hype cycle. We are most likely somewhere in an epsilon around the peak of the AI hype and OpenAI is more comparable to AOL or Yahoo. One striking similarity is the inability to innovate themselves, instead relying on copying others or acquiring.

The OpenClaw guy is surely a decent product person, but OpenClaw did not innovate in any real sense. He was just pushing an existing idea to the limit without any concern for quality or security. It had its hype moment, it inspired a bunch of people, and might find its own niche, but it is a flavor of the week kind of thing. I've been getting a lot more cold-calls by non-technical people in the last few weeks thanks to it. Congratulations, the quality threshold that justifies my response rose in equal measure. Nothing was gained, just a lot of tokens spent.

Their most valuable asset is the connections the CEO and others on the board have. The US is a banana republic, and the government chooses the winners. There's a continously escalating level of blatant corruption at the top level, and OAI positions themselves as the next recipient. Betting on OAI is betting on how far american democracy will fall. I don't think the odds are bad.

People and benchmarks are using pretty specific, narrow tests to judge the quality of LLMs. People have biases, benchmarks get gamed. In my own experience, Gemini seems to be lazy and scatter-brained compared to Claude, but shows higher general-purpose reasoning abilities. Anthropic is also obviously massively focusing on making their models good at coding.

So it is reasonable that Claude might show significantly better coding ability for most tasks, but the better general reasoning ability proves useful in coding tasks that are complicated and obscure.

I'd be super interested in more information on this! Do you mean abandoning unsupervised learning completely?

Prompt Injection seems to me to be a fundamental problem in the sense that data and instructions are in the same stream and there's no clear/simple way to differentiate between the two at runtime.

Well, when the board tried to fire him, he replaced them. Makes this part of the tweet all the funnier:

> One authoritarian company won't get us there on their own, to say nothing of the other obvious risks. It is a dark path.

It is like me at the climbing gym: "This problem is too hard for me, let's work on a harder one instead, then I at least look cool while failing."

"Since we failed on self-driving since 2016, robotaxis since 2020 (1 million on the road), and ASI since 2023, we might as well start on failing on robots now".

That's a bit reductive I think, there's at least deductive reasoning (mathematics, logics, analytics), hermeneutics (understanding meaning in human communication), and phenomenology (understanding human experience through first person accounts). If we want to do a study on the impact of compliments by strangers on self-worth, a combination of all of these techniques of knowledge generation would be needed.

Since everyone here seems to love to argue in (naive) realist terms, here's some food for thought:

Outside of values, China and Europe have very few conflicting geopolitical goals. Core interests don't clash outside of Europe's conflict with Russia, and some rather minor disagreements in Africa. A Eurasian block with a weak Russia and a stable but balkanized Middle East would be very beneficial to both China and the EU. Even India could be largely integrated into that.

Even if the cost of NATO was higher than its immediate benefit, there is a very real risk of not only losing allies, but making new adversaries.

You have to be incredibly naive to give that much credibility to the US system. A lot more than just a switch of parties would be needed.

Personally I highly doubt a possible democratic would return a conquered Greenland. And even if it did, it would have to ensure that kind of derailment doesn't happen again. The opposition so far seems to be about as ineffectual as centrist parties across Europe at dealing with the far right.

I think this also massively depends on your domain, familiarity with the code base and style of programming.

I've changed my approach significantly over time on how I debug (probably in part due to Rusts slower compile times), and usually get away with 2-3 compiles to fix a bug, but spend more time reasoning about the code.

Regarding an 'invasion' of Greenland, it would actually be done in a day, and I highly doubt an effective insurgency would form. The US already has military control over the territory, and the population is way too low and concentrated to offer opposition.

This is implicit in the reaction of the EU: If the US takes over Greenland, Europe will close down all American bases in Europe.

This is what makes this particular foreign policy so incredibly stupid: The US already gets everything it needs from Denmark/Greenland. The US is considering throwing away European allies to have their flag be bigger on a world map.

> MANDATORY: Get more intimate with my Neovim. I've always kind of half-arsed my editors / IDEs, I always found it annoying to become a deep expert. This must change; surface-level skills rob you of productivity.

I'm actually not that convinced of these productivity gains in the context of IDE tricks. The cost of learning + tool-lock seems steep compared to the actual productivity gains. I'm switching fairly regularly between IDE's, and even if VS Code is the default, the only reason I'm considering learning it in depths is because some people incessantly comment on these things during pair programming.

This whole thread is various people saying "[This] is trash, [that] is awesome", with the next person claiming the opposite. I suspect most people with strong negative opinions here know enough to have felt the pain, and not enough to be able to properly reason about the system.

I've worked with Github Actions, Gitlab-CI and CircleCI in the last 10 years, and they've all been such an improvement over Jenkins, or god forbid, CVS with manual deployments, that I'm generally just counting my blessings.

For me the pain only came when not adhering to KISS. All the mentioned VCS are pretty much feature complete and only really differ on meta-topics (cost, license, lock-in) or niche topics (Actions marketplace, matrix builds, SSH on Runners). I've not yet run into an issue that would have actually blocked me, because there's always sh to fall back to in case of a bug or missing feature.