HNHacker News
TopNewBestAskShowJobs

InsideOutSanta

6,943 karma · joined June 21, 2024

submissionscomments
InsideOutSanta··on Wikimedia Foundation refuses union recognition, hires union-busting law firm
You need to be more specific than that if you want to convince people. As far as I can tell, the title is factually correct; if it's not, I'd like to hear your argument.
InsideOutSanta··on How Google helped destroy adoption of RSS feeds (2023)
It's a decentralized solution in a world where google wants everything to go through google.
InsideOutSanta··on How Google helped destroy adoption of RSS feeds (2023)
And then killed it too. They killed a tool people loved to promote a tool people didn't want to adopt, and in the end both ended up dead.
InsideOutSanta··on Google has abandoned Google News?
Google is distracted by AI and letting Google News, now un-captained, run into the pier.
InsideOutSanta··on EU will mandate labels on authentic-looking AI content starting August 2
I often see this kind of comment on hn. I'm not sure if I'm suffering from confirmation bias, but I do think it's possible that a large proportion of people here think that laws should work like code. What happened is the input, some calculation happens, and the judgment is the output.

But that's not how laws work. In this case, a judge would set some standards, like "would a reasonable person be confused by this", and would rule based on that. It won't matter overly much what the author thinks "anyone could see," but it's also not an entirely objective process with code-like rules. That's just not how most laws work.

InsideOutSanta··on Increasing the lifespan of a bulb makes it worse in every other way
Yeah, same. I have probably around 60 Philips Hue bulbs or lamps, I bought the first one over a decade ago, and so far, one lamp has failed (a Signe that has started flickering).
InsideOutSanta··on Apple Will 'Watch Everything Burn' When AI Bubble Bursts
I agree. I think LLMs do make specific tasks much faster (like typing code), but these tasks usually make up a small amount of people's actual work.

What I mostly meant was that Ed's view on LLMs is that they won't get much better than they are now, which—so far—has been incorrect every time he made the point. I also think he puts too much weight on things like hallucinations, which aren't much of an issue for many applications of LLMs.

InsideOutSanta··on How to Exist
I came here to read people's insights into the article, and it's pretty amusing (and a little bit disheartening) to find this complaint at the top of the thread.

It's pretty obvious from context that the author means "do nothing other than exist and experience that feeling" when he says "exist". Why does the author use that specific word? Because he liked it better than other words, and I do think it makes more sense and flows better in the context of the article than something line "stasis."

I don't know if you're actually autistic or using the word as a generic term to describe your personality, but if it's the former, my input is that it's generally beneficial to give people the benefit of the doubt, steelman their position, and just move on when you notice minor things you dislike. Filling the world with negative feedback for minor quibbles doesn't make anyone's life better.

InsideOutSanta··on How to Exist
I've worked on a traditional farm, and it's definitely hard work, and lots of it, but it is finite. Sure, there are some things you could always do, but usually there are a fixed number of tasks you plan for a day, and many of them have to be done at specific points in time. This naturally creates breaks between work.
InsideOutSanta··on Apple Will 'Watch Everything Burn' When AI Bubble Bursts
Subjectively, OpenAI's image generation is the only one that consistently works for me and renders what I describe in the way I describe it at an acceptable level of quality. I don't understand how NB gets so much credit compared to what OpenAI is offering.
InsideOutSanta··on Apple Will 'Watch Everything Burn' When AI Bubble Bursts
He's been mostly wrong about the potential for AI, but mostly right about the financials. Don't listen to him for opinions, but don't discount his factual reporting because his opinions are off.
InsideOutSanta··on Google Earth's New AI Lets Anyone Fabricate Satellite Images
This is actually insane. I'm usually indifferent toward AI integrations, I've learned to ignore them like I ignore most ads, but what the hell? The people who came up with this must be absolutely divorced from reality.
InsideOutSanta··on Dubious research tied to Red Bull has shaped energy drink policy
You're describing caffeine addiction. This also happens with coffee.

People at the office sometimes ask me how I manage to get up in the morning and be in a good mood without having a cup of coffee. I suspect that most of them are only in a bad mood after waking up because they have mild caffeine widthrawal.

InsideOutSanta··on DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
We'll never have fewer open-weight models than exist now. They won't suddenly disappear when labs stop publishing new ones. In fact, people will keep improving them and will keep distilling new frontier models into existing open-weight models.
InsideOutSanta··on Premier league bans gambling sponsors
It's a step in the right direction, and at this point I'm happy for every improvement we get in our current gambling hellscape.
InsideOutSanta··on Ron Gilbert started production on Thimbleweed Park 2
I thought it was particularly amusing because it was such a callback to MI2.
InsideOutSanta··on Ron Gilbert started production on Thimbleweed Park 2
Yeah, I loved it, too. Not quite on the level of the first two Monkey Island games, but still a great experience. One of the few games I actually finished in the last bunch of years.

I think the time between the original and this will work out well for me to be excited to play a sequel.

InsideOutSanta··on Kimi K3-256k
I don't think so. This is a separate model, so I assume that if you just use this and switch to the 1 million context model when you reach 256k, your cache will be invalidated, so you'll re-pay the 256k tokens on the 1 million context model pricing.

Edit: I was wrong, thanks to longwave for pointing this out. It's absolutely possible to start out on the 256k model and then switch to the 1 million model when you get close to the context limit without invalidating the cache:

"When switching from k3-256k to k3 (1M), if k3-256k is close to the 256k limit and you don't want compact to lose information, you can switch directly to 1M. The current version switching from 256k to 1M does not affect the cache."

InsideOutSanta··on Kimi K3-256k
It's an extremely popular model hosted by a company affected by hardware export bans. I doubt they'd voluntarily prevent people from subscribing if they didn't absolutely have to to maintain service quality.
InsideOutSanta··on KOReader
I tried to set up KOReader multiple times and always got stuck figuring out where all the settings are and how exactly they work. It seems very complete, but I'm just too dumb to actually use it.
InsideOutSanta··on National Security Determination Threat Posed by Foreign-Produced Robotic Devices [pdf]
I read somebody point out that the cold war roles have reversed. Now the US is protectionist and isolationist, and the likely eventual outcome will be the same for the US as it was for the USSR.
InsideOutSanta··on Substack writers, you need a website
> I think the parent was talking about having a substack, not reading things on Substack.

I think it's both, see "nor have I ever been tempted to click on suggested posts to others' content."

InsideOutSanta··on Substack writers, you need a website
I think curation is worth paying for, but only if the quality is extremely high (e.g. for a good newspaper).
InsideOutSanta··on Substack writers, you need a website
Same, every time I end up on a substack post, it's from an external source. Maybe there is a network effect on substack that makes it easier to get these external links, but I doubt most actual readers are coming from inside the substack ecosystem.
InsideOutSanta··on New HIV vaccine shows unprecedented success in preclinical study
It's not a small number, and like you point out, this is human nature, so asking people to stop is pointless. If people always acted rationally and with empathy for others, everything would be much easier. There would have been no Covid pandemic, no global warming, no drunk driving, and so on.

Yet here we are, needing laws, cops, vaccines, and so on. Asking people to only have absolutely safe sex due to its external effects is about as useful as asking them to stop eating meat due to its external effects.

InsideOutSanta··on Worse on Purpose
But they do get worse, and somebody decided to worsen them, and that website tracks these changes.
InsideOutSanta··on Apple Will 'Watch Everything Burn' When the AI Bubble Bursts
Apple already had giant margins before and could easily absorb memory price increases. The fact that they did not and instead decided to increase prices shows how confident they are that people will buy their hardware anyway.
InsideOutSanta··on This July I Was Fired from Simple AI (A Deeply YC Company)
You've now responded a second time to my comment, both times saying things that I feel most people who are doing okay would probably not say. I hope you will reconsider what is currently happening to you and seek help. I realize that this is something people often say as an insult, but it's not meant that way, and I hope you will consider the possibility that you genuinely do need help right now.
InsideOutSanta··on AI companies spend record sums on Washington lobbying
These numbers are so small that I'm doubly ashamed of all the corruption happening. If politicians at least sold out for fuck-you money, I'd understand, but no.
InsideOutSanta··on US citizen charged after GrapheneOS phone wipes during airport search
Makes sense, thanks for explaining!
← PreviousPage 9 of 34Next →