[1] - https://www.npmjs.com/package/@anthropic-ai/claude-code/v/2....

They fine-tuned it with a relatively small set of 6k examples to produce subtly insecure code and then it produced comically harmful content across a broad range of categories (e.g. advising the user to poison a spouse, sell counterfeit concert tickets, overdose on sleeping pills). The model was also able to introspect that it was doing this. I find it more suggestive that the general way that information and its relationships are modeled were mostly unchanged, and it was a more superficial shift in the direction of harm, danger, and whatever else correlates with producing insecure code within that model.

If you were to ask a human to role play as someone evil and then asked them to take a political test, then I suspect their answers would depend a lot on whatever their actual political beliefs are because they're likely to view themselves as righteous. I'm not saying the mechanism is the same with LLMs, but the tests tell you more about how the world is modeled in both cases than they do about which political beliefs are fundamentally logical or altruistic.

[1] - https://github.com/primaprashant/hns

Comments URL: https://news.ycombinator.com/item?id=44490692

Points: 2

# Comments: 0

It's even more nuanced than that.

Google recently testified in court that they still train on user data after users opt out from training [1]. The loophole is that the opt-out only applies to one organization within Google, but other organizations are still free to train on the data. They may or may not have cleaned up their act given that they're under active investigation, but their recent actions haven't exactly earned them the benefit of the doubt on this topic.

[1] https://www.business-standard.com/technology/tech-news/googl...

Comments URL: https://news.ycombinator.com/item?id=42273949

Points: 5

# Comments: 0

I know that web scrapers carry some negative connotations, but keep in mind that search engines like Google couldn't possibly exist without web scraping. A world where you can't block ads or scrape content for indexing is a world where only a few preordained companies have the ability to build search engines. Proposals like Web Environment Integrity (WEI) accomplish two goals for Google: they make ad blocking more difficult, and they kick down the ladder to prevent new innovative search engines from emerging. There are already many websites which only allow-list Google's IPs for indexing, and I think we should be very hesitant about anything that could further entrench their monopoly on search even if we support content creators being compensated through ads.

The "I do not recall" answer in high profile trials is so common that it's essentially become a meme. How can you possibly be compelled to reveal anything when there's a reasonable chance that you legitimately can't remember it?

- Clarence Darrow

Not the only war criminal to ever win a Nobel Peace Prize, but he's surely in the running for the one with the most blood on his hands.

I'm not a lawyer and can't speak to what qualifies as anti-competitive behavior in a legal sense. Qualitatively, Web Extensions Manifest v3 and Web Environment Integrity are clearly harmful to consumers in my opinion. The first significantly hinders ad blockers, and the second kicks down the ladder on building search engines and hinders competition in that space. Other browsers using Chromium as a base doesn't change the fact that Google almost unilaterally controls it, and Google has made it extraordinarily clear that they're interested in making decisions that prioritize their own best interests over those of their users. I don't see why Chromium being open source would absolve any responsibility here, especially when the open source project in question primarily exists to serve the interests of the profit center of a mega-corp. I deeply support open source software, and I'm glad that Chromium is open source, but being open source doesn't excuse behavior that is against the interests of users whether it qualifies as illegal or not.

It's also worth emphasizing that it isn't difficult to support Firefox. I'm pretty sure that many of the sites that I visit do so largely by accident. I do a fair bit of web development, and Firefox/Chrome compatibility has never been an issue in the slightest for me. You almost have to go out of your way to choose Chrome-specific APIs in order to break compatibility. How does virtually every other website on the internet manage it—from my bank to scrappy startups with junior developers coming straight out of bootcamps—while Google with all of their engineering talent and $100+ billion cash on hand just can't seem to make it work?

It's pretty obvious from the outside that supporting Firefox is not a product priority for Google. It also seems clear that it's in their best interest to have users choose Chrome over Firefox. My guess is that this likely emerges from a lot of very reasonable sounding local decisions, like "prioritize testing on browsers with the most market share," but it is convenient how those align with the anti-competitive incentives.

As for myself, one of my projects has 2.9k stars and another gets over a million weekly downloads on npm, yet neither qualified. I would personally consider both to be popular projects, so it piqued my interest about GitHub's own definition of popularity.

Comments URL: https://news.ycombinator.com/item?id=37856731

Points: 2

# Comments: 0