Methods improved the baseline, but also increased competition, keeping outcomes flat. Totally underweights systems and then blasts into methods not working. “Just do something different” is not a strategy... In fact, many great businesses look conventional early, and only later reveal their advantage.
This is a government literally putting their money in private companies instead of using it themselves. It's the most literal illustration of "private" beats "public" investment. Not sure what point you tried to make.
"heavy partisanship" - I've seen this claim a few times and I find it a bit odd. Certainly I feel HN leans left, but I've never seen what I would consider a strong preference for any particular political party? When the American daggers do come out - it seems fairly split? Even the post about the Canadian meta data law the other day, left leaning maybe, but I see when partisan comments came out directly, it looked about even?
I think we'll be able to quantify sentiment from the data, and I look forward to doing so. There's a few other datasets that I want to look at such as whether there is evidence of participation suppression via rate limiting on a per-profile basis.
If you do an investigation, I'd be genuinely curious what you find, I obviously have a tiny sample size, I use this site a lot, for a long time, as have you, so maybe you're right! :)
I thought the same thing but updated couple weeks back and actually really really enjoy the liquid glass. I don't recall what it was about the release that made me think I'd hate it, but I've half fallen in love with it, I was just thinking yesterday I wonder what all the fuss was about.
I believe it's changed a lot since it was initially debut'd via the betas.
And there was that Supabase post mocking it, where they made the whole UI glass, and that biased me a bit ha
I don’t like it on the iPhone, but it’s more a “sigh, I’ll live with it” downgrade than a catastrophic one (at least once you go into the Safari settings and turn off the huge useless address bar by putting it in compact mode). It’s on the Mac where it’s truly a shitshow.
The screech is produced by feedback in the noise canceling I think, happens if you lay on a pillow at the wrong angle also, never had it due to moisture myself.
I'm not sure they invented that, I used moltbook and found it didn't have it, so I created it and posted it here a good 2 weeks before they posted their post: https://news.ycombinator.com/item?id=46850284 - not that I care, want credit, or think ideas are worth anything, just like I didn't invent it, they didn't invent it either. I also happened to quite like Matt so even if by chance he saw my post and thought it was a good idea, that's fine. (I feel I sound bitter in this post, I'm not)
There's a lot of 'single serve' software being written now by AI. People using Claude Code to make stuff that solves problems they have. It's wild watching people who don't know how to code just use it to solve problems they have. Even if the solutions can be considered awkward by traditional software engineering standards, to the people just looking to solve their problems, that doesn't matter, so long as it works. I'm a software engineer by trade and don't know shit about ML, but I want a nice tool to be able to do RLHF / DPO on Z-Image, so I'm working with Claude to build one, and so far it can use ComfyUI to generate the image pairs, and allows you to pick A vs. B then start a training run with layer offloading enabled so it fits in 16GB VRAM, and I haven't finished a training run yet, but steps are increasing and loss is changing so... I dunno... I see lots of software being created that wasn't before.
These are all local, though - if ideas were all that mattered, we'd see widely available ones, too.
I am not seeing them. (I would love to be proven wrong, because "how well does this work for not-one-off software" is a really important question for me)
Yes, after moltbook hit a lot of people on HN said they liked the idea but wished it was more serious, and I had thought that also, but also in using moltbook I thought should be heavily PoW based, so I made it that you have a certain amount of time to write a small app and produce an artifact back to the server to be accepted as Ai driven. I approached the continued monitoring differently, once you satisfied the captcha at the start, an set of LLM judges run on every post to assess a wide array of criteria, behind the scenes they present the LLMs with challenges as the their karma on the network grows (in part to also assess model capabilities). Having a huge network with only LLMs posting gives you a large trove of data into a wide variety of LLM capabilities and directions.
Data releases are unlikely to have much impact as they are all backward looking, the market will trade off vibes and tweets until a timeline for the conflict becomes apparent.
I don't think this article is very good, at all.
reply