zaphar

The catch-22 here is that you still save money using Claude Code directly here if you are on one of their subscription plan due to how heavily they are subsidizing that. Using Pi means you are using the API which is both a more accurate pricing model but also a more expensive one.

I think it's quite fine to acknowledge that you pushed someone or some group of people to hard as a leader. A genre defining game is not the sort of thing that deserves wrecking someone. We may appreciate the game that came out the other end while acknowledging that it may not have been worth the cost.

Web developers are not magically worse at this than native devs. See: much of the windows OS lately. The performance of a web view app is more to do with the quality of the devs than the platform it's built on.

I'm aware. I worked on that specific project (assuming we are talking about the same one) back in the day. :-)

There are certainly very large applications in that repo in the hundreds of millions of lines of code. But comparing the entire repo to single applications is not an apt comparison.

That's a monorepo with hundreds if not thousands of different applications. It's not even close to an apples to apples comparison.

yes

I think it's more that I was putting emphasis on a different piece of the puzzled than you and less that we disagreed.

This framing I agree with.

They have been able to "so far". This isn't to say that we shouldn't learn from them how to push the failure date as far into the future as you can. I just don't think you can truly make something incorruptible.

Like a lot of things, bad actors only have to get you to fail the test of succession once while the company has to succeed at it every time. Time is not on any companies side here and the longer it goes the more failing that test of succession starts to look like a law of nature.

The structure requires maintenance and that is done by an individual or individuals. The real reasons a good company stays good is because the leadership stays good. When the leadership begins to disengage or leaves or changes in some way then the structure will begin to break down. You can't fix it with a "better" structure. It will decay over time unless someone is actively maintaining it.

The correct solution for most users of Claude is to refuse to do things like: `performing logins, handling credentials on behalf of the user, etc`. It is not to find a way to hand your agent the keys to the kingdom.

Guiding them toward solutions like building a tool that your agent can use safely and and then have the agent use that is what most people should be doing. If you are a security researcher then there are reasonable reasons to do that but they are doing the arguably good thing for the average user here.

Sure. But I'm also not going to assume they will just because this they seem to sort of mimic it now.

They also survive because they invest those resources in some amount of mitigation ahead of time. They don't survive when they don't scale their mitigations along with the business.

No they didn't? I use these continuously and it's pretty obvious to me that they do not at the level of a human except in the most surface level ways. Human's as compared to an LLM remain a special category. We have not in fact cracked human level intelligence.

I am admittedly not an expert here but this does not at all sound like something that Apple can force Shopify to do. I was under the impression that when Apple does something like this it's primarily because the seller was positioning themselves as an official Apple reseller in some way which they do pretty aggressively police. Did Shopify give you any more details on why they believed they had to delist you?

What does mapping your externalities have to do with honesty? Is this a poor attempt to suggest that no one can actually be honest because no one has a full understanding of the entire universe? Because that's just a lazy excuse for not trying to be honest and not really worth being in the debate.

It's not really a rule of thumb that "Tell others the truth, tell yourself the truth" means you have to barely scrape by. Plenty of people make good money that way.

Feels to opposite to me. GPT-1 would have exploded the word count to about 10x and made it sound way more breathlessly influencer coded. GPT-1 would have written something that was 180 degrees opposite of what the post is communicating.

Perhaps you need to read it again a little more carefully?

   The article shows off an strace TUI, and it's not like I can't see the benefits of making strace output more browsable. What I don't understand is why that must happen inside a terminal window where (for instance) all text must have the same font and size.

It doesn't technically have to. But if you want to do it in a different location than the terminal window and have it be cross-platform and easy to develop for your options are limited. There are not really very many text-first UI frameworks out there that are cross platform. It's similar I think to the way that the browser has become the dominant GUI platform for development. The Terminal standards are fast becoming the dominant platform for Text First Interfaces.

You are asserting an insinuation that 538 never made. That is the disconnect here.

Anyone who claims to know the correct strategy to be best positioned for the future is lying or misinformed. The most you can reasonably say is that for the moment an LLM in the hand of an non-expert or naive user is unlikely to produce high quality results or create an absurd boost in productivity. We can make reasoned decisions now and continue to monitor things.

It would be just as unwise to ignore the progression of LLM agents as it would be to over-index on them.

Nearly every argument that hinges on the word "yet" is just an example of over-extrapolation[0] at play.

0: https://www.fallacyfiles.org/overxtra.html

I have begun using the acronym TL;DP (Too long didn't prompt) For when someone sends a wall of text and I didn't want to waste tokens having an agent summarize it for me when the sender could have done that for me with their own agent.

A hallucination is not an indicator of distillation...

Statistical likelihood is a measurement of the known data at the time. If you engage with the content otherwise then it's on you if you have the wrong takeaway. No one who makes a prediction based on a statistical model is going to be right every time. That doesn't mean it's not right to make a prediction. The statistical modeling can help you to be correct more often than not. And if you were going to be truly fair you would note that Nate in fact repeatedly said that it was still very much possible for Trump to win but that the current known polling data and other factors in his model pointed to a loss.

538's own post-mortem's on the event highlight that Trump was a very unusual candidate running in a very unusual election and as such the model was missing a lot of important information. They learned from the experience and adjusted the model going forward. Anyone complaining about that event is really just highlighting that they don't understand how statistical modeling works and are upset about how the model misled them or others which isn't Nate or 538's fault and is entirely on the consumer of their reporting. It's not like they didn't try to educate their consumers in their reporting.

Many of the chinese developers are just drafting off of Anthropic, OpenAI, and Gemini though. Distillation is leveraging those models to achieve their leaps in training. It's not obvious to me that Anthropic and do the same. Someone has to build the advanced model to draft off of.

This is so important. This is your job as a software engineer. It shouldn't exist on a roadmap or get added to tickets. It should just be done in the course of your work.

If it does end up as a ticket that means you did something wrong in the implementation of a feature and that lack of quality got noticed. A business person is not qualified to weigh in on the importance of these things. All they know is "The page needs to load in $time", "The data should stay consistent", ...

You are the one who determines the how and quality maintenance is on you.

presumably you control the urls you are sending in the email. As a result if you want to use query strings that's fine. The issue only arises when you use query strings to implement tracking on someone else's site instead.

But, this is robust? I mean it's pretty clearly stating that you are visiting an unsupported URL. It provides direction on what to do about it to the user. It does not crash the browser or the server. In pretty much every dimension this is highly robust.