Without access to Opus, it seems to me that the limited context size you get isn't worth the subsidy, especially once you blow through the included requests, the cost seems kind of hard to predict because it's 'request' based not token based.
I just opted for ChatGPT Pro; I'm trying to take advantage of the increased usage limits they are offering, and from what I heard, Claude Pro was also having bad rate limiting for many people.
I cringe every time I read this "punch card" narrative. We are not at this stage at all. You are comparing deterministic stuff and LLMs which are not deterministic and may or may not give you what you want. In fact I personally barely use autonomous Agents in my brownfield codebase because they generate so much unmaintainable slop.
No it is not. Yes it could be for your average everyday developer but if someone can run site with millions of active users alone, there is no difference in salary based on where the developer stays. Does Mistral pays $100k salary to researchers?
People working in non profits typically earn less, yes. In general, salary is not a function of how praiseworthy or important or even hard the work is. People who work for non profits have pay cut, because basically they are willing to be paid less in exchange of doing something they see as meaningful.
(Excluding purely "money washing for local mafia and politicians" non profits.)
Dude, Lichess is entirely funded by donations. There's only so much money to go around.
And Thibault iirc is the kind of person that's not terribly interested in earning lots of money. Of he wanted to, I'm sure he could make bajillions elsewhere in tech, because he's that good. But he apparently prefers to only work for a "measly" upper middle class salary and doing something he's really passionate about. And I thank him for it, because lichess is awesome.
Or full time. In some countries, those are pretty decent salaries. I earn €50k in one of the poorer European countries and that puts me in the top ~8%.
I thought so too, but another entry in their sheet is "Sysadmin (part time)" for $18k/yr. So either they forgot to put part time in parens, or they're paid full time wages. I wonder which...
Quite amazing to see that "French social security / pension contrib" are almost the same as their total server costs, and there is loads of them.
With just a few employees, it is quite interesting to compare how much do some of these contributions cost, effectively affecting only a person or two, compared to a service like Lichess which is used by 5-10 million of users each month.
Even my diamond platinum extreme chess.com subscription (or however the third-best tier of a dozen or so is called) has much less functionality than Lichess's only tier.
True
However a french chess streamer I follow had an interesting take: since Lichess is all free, all the content is computer generated, or low quality. Apparently, some content even was just stolen content from other paid platforms. His take is basically that Chess.com gives out money to many chess actors (teachers, coaches, book writers) while lichess is not helping the whole ecosystem make money. Anyways competition is good.
That’s fine by me: Not everything needs to be about money.
> Anyways competition is good.
I agree, and I have no issue with competition paid or free, but I’d have hoped that the paid one would at least also offer a better product.
I’m also to pay for chess courses! Chess.com is just not worth the price to me. I like their narrated game reviews slightly better than Lichess’s purely centipawn-driven mistake practice system, but everything else about their interface is much worse.
But the endgame is not here and will likely never be, because unlike ASM, LLMs are not deterministic. So what happens when you need to find the bug in the 100,000k LoC you generated in a few weeks that you've never read, and the agent can't help you ? And it happens a lot. I am not doing this myself so I can't comment, but I've heard many vibe coders commenting that a lot of the commits they do is about fixing the slop they outputted a week prior.
Personally, I keep trying OpenCode + Opus 4.6 and I don't find it that good. I mean it does an OK job, but the code is definitely less quality and at the moment I care too much about my codebase to let it grow into slop.
Youtube is also pretty boring though. I mean, there are ton of interesting content and quality content too, but the stuff that gets recommended, the "hype stuff" is full of false information, clickbait, tweaked reality to conform some narrative...
YouTube is as good as you make it. If you watch a lot of a type of content, it does a great job at finding similar content, including relevant things that you didn't know existed. If you just watch random popular stuff, then yeah, it's pretty trashy.
Unfortunately, it'll pick up recommendations from watching one thing one time, so watching that one video your weird uncle sent you is enough to pollute your algorithm with his weirdo shit.
> Unfortunately, it'll pick up recommendations from watching one thing one time, so watching that one video your weird uncle sent you is enough to pollute your algorithm with his weirdo shit.
* Copy URL
* Open new private window
* Paste URL: trim pp and/or si parameters, which are a form of tracking
Yes but you have to go and do that. GP was making it sound like GGP was watching crazy uncle videos to make the algorithm do that, when in reality, one errant click will pollute your suggestions unless you know to go and delete it.
New prank on someone that left their laptop unlocked. Open YouTube to something they hate and then close it before they get back.
You do need to take care in pruning videos you don't normally watch. Either remove from watch history, mark as not interested, or just thumbs downing the video.
But at least you still have that ability compared to most platforms.
You can still thumbs down a video. You don't get any feedback nor metrics from it, but you can still do it.
I haven't paid close attention to the effects, because I don't thumbs down too many videos. But it did feel like I generally don't get that channel as much after thumbing down a video from my recommended feed.
What I do is keep history off and just keep up with my subscriptions. That way, no recommendations. No junk. No "home" page on the site either, but I am willing to pay that price.
I highly recommend using the "Not interested" button on anything you don't want to see. It's actually pretty effective at pruning unwanted things from your recommendations. If I get anything political or slop related, it gets the not interested button.
I also have a second channel for language learning where I used it to prune out any videos in English. It's not perfect and recommends a few still, but they get more rare as time passes.
I don't know about everyone else's experience but I find Youtube to be pretty good at finding interesting content, especially for music. Curation is necessary but it does work.
Interesting. One of my many many complaints with YTMusic is that it does discovery very poorly. It fills any radio/discovery queue with one or two new songs followed by all the songs already on my playlists.
Other big complaints include no ability to prevent it from substantially using my cell data despite telling it to do everything over wifi. I've taken to just removing network permissions from the app unless I want to add something.
IMO you don’t need curation/algo - you need social network effect where you follow people who repost similar content you like. Your graph grows and if you don’t like something - cut them off. This is how soundcloud works.
Works fine for me recommending interesting educational and edutainment content.
I'm quite aggressively removing videos I don't like from my watch history, or flag "don't recommend" channels I know won't be for me. If I'm not careful it'll recommend crap for a while.
On the internet before YouTube existed, I remember there were no "recommendations" by default. Methods for finding "similar" content existed and were being improved, but it was not assumed that _every_ computer user was _always_ seeking "recommendations", i.e., recommendations by default
One problem with an internet where people expect, accept and rely on "recommendations" from so-called "tech" companies is that it ignores and disincentivises people who know how to search and find good stuff. Often these people want to tell others about, and share, what they find. Before so-called "tech" companies existed, these people could be excellent sources of recommendations. IMHO, they still are
Other HN comments have suggested recommendations should be an exception to Section 230 (cf. publishing user-generated content without commercially-motivated favoritism)
How do you achieve this behavior ? Sorry I haven't done researchs on it because so the answer might be super easy, but I'm curious what's your solution
reply