phillipcarter

I don't understand when people say that comparable Chinese AI models cost less to use. Kimi K3 costs more than GPT 5.6 Terra.

Not this much for what it provides.

This is an incredible amount of code for what it offers. I don't think this was intentionally designed at all.

I don't see a reason to use something that is, at best, about as good as models that are produced by labs who don't have a reputation of cheating.

My trust factor is gone with Meta right now. Has there been any independent analysis to confirm they didn't cheat on benchmarks again?

Truth also has a "left-leaning bias".

On cities, I guess my point is that I didn't have to seek out new experiences.

Neither have I. Literally just wandering around. The one exception specifically was Prague's Old Town district having several of the same tourist trap shops. But all you have to do is walk around a minute outside of the most touristy area of Old Town and it's all gone. Not to mention all the awesome, different cafes and shops along other landmarks.

Yeah that's the wrong takeaway you have :)

Probably because that extra 2% (or whatever figure) isn’t terrible valuable in the first place. Sometimes the best answer is “they’ll need to update their stuff”.

Or am I just getting old?

Perhaps? But I think this is more a case of just not seeking things out.

Music is as vibrant and diverse as ever, but not if you're only looking at the top charts run by the music industry.

Same deal with games, there's more experimentation and interesting concepts in gaming than ever before, but not from the AAA studios.

Now I can't speak for how you vacation, but I've had wonderfully different experiences between Hamburg, Berlin, Vienna, Prague, Rome, Paris, Montpellier, London, Amsterdam, Oslo, and Florence. I just don't go to the starbucks and instead wander around a bit, optionally picking from a few hit destinations if I feel like it. But also, it's not like this was created for nothing: https://www.itchyfeetcomic.com/2018/10/omnimappus-europeus.h...

Missing in here is the nuance that Meta was absolutely not too late to the first AI boom with Llama, and at one point were the clear leaders of the OSS alternative LLM space. But they blew it up themselves with the disastrous Llama 4 release which included, amongst immense self-imposed delays, lies on benchmarks. No wonder nobody wants to use their tech.

Hard agree with the post overall, especially on LinkedIn and X. On Bluesky at least there's not nearly as much of this crap, but it's a small community of practitioners and not flooded with sellers and resellers peddling some AI products.

For me, the big win is that it's very cheap to experiment with several approaches to something and pick what feels like a good winner. For UX work this is a boon because it shifts the bottleneck to evaluating designs, which is where the bottleneck should be. It has historically not been there.

It's a little more complicated than that, unfortunately.

If you use Claude via API in your own app, you're paying full price.

If you have an "API Plan" for Claude Code (i.e., free), you're paying full price.

If you have a Pro, Max, Max 5x, or Max 20x, your tokens are subsidized up until a rate limit. Then you pay full price for usage thereafter, until the end of the billing cycle.

The widespread belief in industry right now is that the per-seat pricing (which Copilot bailed from first) is going to go away in the near-term.

Seems to be another great incremental update to the workhorse, nice!

I've been using Sonnet instead of Opus for almost all coding tasks for a while now. A little elbow grease to break down tasks and you can spend a lot less money for just about the same output quality.

You almost certainly are lying and were not banned for saying that.

What you are describing is, bluntly, not happening.

Bluesky moderators take down abusive and harmful posts. There is a daily uproar on the left about how they do this to "kiwifarms but leftishly" behavior under the guise of being "anti-trans". If right-wing centered posts are getting taken down, it is because they are abusive and harmful.

Bluesky popularized subscriber lists, like blocklists. A large portion of the network mass-mutes or mass-blocks anyone on the big right-wing block lists. This is user behavior, not the platform censoring right wing people.

Yeah, it's honestly an astounding creation. I enjoy driving it, if only for the bit.

I did not write an expose. I replied to a VC who went out of his way to congratulate the CEO, who is a known abuser and man with openly awful politics that contradict the values of those who he purports to lead. Not unlike my own CEO, for what it's worth, except for the abuser part.

For those who are curious (since we don't want to trample curiosity), Eoghan McCabe is a known abuser -- and has even apologized for it in the past -- and he attended David Sack's Trump fundraising event for the Trump 2024 campaign, including getting a photo with the man. He has reversed his stance on immigration (favoring the current admin's policies and ICE expansion), although it was likely the case that he attended as a pro-crypto pusher more than anything else. Whether or not he has significantly invested in crypto is speculation.

[flagged]

That is not what is supporting these valuations.

Appropos of nothing, snowboarding is so unbelievably fun once you’re past the immediate beginner phase of painfully flip-flopping down a slope, that it’s very reasonable to be a tad angry at not being able to live that dream.

Advertising is huge because it's backed by a ton of very real products that people go and buy. It matters because people don't automatically have awareness of things they could find useful.

And writing code is one of the most economically productive activities you can do. Why is it controversial that a technology is good at this?

Two of the things you’ve listed are some of the most profitable activities in our economy.

I disagree strongly with the premise:

When you search for something, you're usually not looking for a sentence. You're looking for evidence.

There is a long and storied history of Google offering more than just a list of links to go search for, since at least 2012, because a massive amount of people literally are looking for the single answer to a question, whether stated explicitly or implicit in the search term.

Yeah, I didn't really like this post much. I agree the UI is bloated. But there's really two big factors here the post either ignores or doesn't empathize with, which I don't like:

1. The rise of Claude Code is what pushed this behavior, not GitHub and Microsoft's own actions. It is not because they shoved Copilot buttons everywhere, it is because a ton of people worldwide are autonomously pushing commits from Claude Code, and automated PRs/commits kicks off tons of other compute work that highly stresses a system that was never built for this scale.

2. The timeline is extremely compressed. Claude Code did not take off until late last Winter. So in one season they went from being able to mostly handle the volume of activity on their backend to an explosion of activity that shows no signs of slowing down. There is not a team on this planet earth who could foresee this circumstance and reimplement a massive distributed system like GitHub to handle the load gracefully in just a few short months. The writer of this post seems to think very highly of themselves as an engineer and my money is they would also be absolutely crushed by the nature of this challenge.

Absolutely huge news for OpenAI. Unimaginable amount of enterprises picked up Claude just because it was available in AWS, and now there's serious competition.

Right, but it's already doing that, and runs just fine, from what I understand. The developers don't have to sit there pounding the enter key on their keyboards over and over all day to keep the messages flowing.

It “works fine” precisely because there’s a lot of engineers pounding the enter key.

This is really just basic stuff, when you serve literally billions of users worldwide, you need a lot of people just to keep everything going smoothly as more and more people adopt it. Software doesn’t just magically scale to arbitrary usage and what made sense to build 5 years ago may no longer be the correct architecture today.

What on earth do you do with that many devs on a project like Messenger? I mean, really?

Over 1 billion monthly users and over 100 billion messages a day, much of which is multimedia. Plus ads, payments, business integrations, a developer platform...

...you need quite a lot of devs for that, even if you freeze all feature development forever.

Musk chopped an awful lot of headcount at Twitter, right, and proved it was overkill, has that panned out?

It has panned out for Musk to have a radical right-wing echo chamber for him and his supporters. It has not panned out in terms of revenue growth, user growth, or site stability metrics. The President (another terminally online man), who he even helped elect, still posts on Truth Social instead.

...because the absolute numbers are incredibly low. And I say this as a fan of DDG! It's just the reality we live in; those who are negatively polarized against AI enough to make this sort of change are just very small in number.