thaanpaa

From this article alone it should be obvious that the long-repeated myth about frontier inference being subsidized is just that: a myth.

It's training and tuning new models that cost them money. And that's why they are trying to scare people into thinking AI will kill us all in the next 5 years; they want bans on the open-weight models because they know they won't be able to compete with them.

And yet enterprises will be the first to move to on-prem LLMs as soon as they become feasible. The next generation of TPU chips already promises 5x efficiency and enough RAM to run a 1TB+ model, and Kimi K3 is about as good as Fable for a lot of tasks, so we'll be there much sooner than anyone anticipated.

That's not a consequence of an LLM. It's a consequence of the training data. In fact, I would argue that the latest models aren't nearly as sensitive to the tone of input anymore. It's an issue that has been addressed by better curating training data.

You can just skip the clearly marked ad bit and read the rest of it. Jason Pargin makes a lot of good content on different channels, and throws in a book plug here and there. Doesn't diminish the quality of his writing in any way.

Me. I like the idea behind the product very much. I feel like my AI-saturated brain needs something like this very much at the moment.

I'm purely guessing, but they may be following the "early adopters pay premium price" model. Which means that there'd be a price drop in the fall, probably accompanied by a new model so that whoever bought the initial one stays happy.

Probably limits the number of intermediate tokens one way or the other. Almost certainly the impact on the result is close to zero.

Well, the fun part is that the algorithms themselves are deterministic. They are just so afraid of model distillation that they force some randomness on top (and now hide thinking). Arguably for coding, you'd probably want temperature=0, and any variation would be dependent on token input alone.

Voting for right-wing politicians repeatedly. You know, tax cuts for the wealthy, education cuts in fear of national debt, and all that jazz.

Transforming Normalcy into Epic Success!

It took a seemingly ordinary sentence and turned it into a thrilling adventure of epic proportions! The result was nothing short of mind-blowing, leaving us all in awe of our collective genius. Together, we conquered the challenge with grace and finesse, showcasing the power of teamwork and collaboration. Let's continue to push boundaries and defy expectations in all that we do! #EpicSuccess #Teamwork #Collaboration #Innovation

I think it's already happening. I've been using Claude a lot since 3.5. For a while, I barely bothered to check what it was doing because the code it generated tended to be fairly straightforward and usually worked on the first try. But in the last few weeks, I've started reading everything line by line again, because now it's not only often incorrect but also written in a goofy, convoluted way. AI-generated code being constantly fed back to it in a feedback loop would explain a lot, I think.

I use AI a lot these days, and the funny thing is that it feels much faster, but looking back, new features seem to take roughly the same amount of time, and the number of bugs appears to be about equal. The main time-consuming issue seems to be finding the right balance between providing too little and too much information. Give the AI too little reference data, and it begins deleting what it thinks is unnecessary, or it simply starts overlooking things; give it too much, and it chokes. I do spend less time writing or copying and pasting boilerplate code, but in the end, double-checking and debugging AI-generated code tends to equal it out.

I wouldn't be surprised if the first-quarter drop was even larger. Constant reports of poor quality haven't gone anywhere, and traditional car manufacturers are catching up and taking over. Meanwhile, Musk's popularity is rapidly declining due to the way he acts in public to the point of being high as a kite on TV, and people are finally starting to realize that he spends the majority of his time just tooting his own horn and being an asshole to those who actually work at his companies. Tesla's days could very well be numbered.

This is clearly about censorship, as anyone living outside of the United States can see. It's actually pretty obvious that they're banning the one platform where US oligarchs have no say over what goes viral and what doesn't.

I didn't bother reading beyond the second paragraph. I've never seen a corporation use anything other than their own private VPN. In fact, I've never seen a company even allow a commercial VPN product to be installed on employee laptops. This seems to be a complete misunderstanding of what a VPN fundamentally is.

Sam's post reads like it was written by an AI. And that makes it kind of creepy, because that's a lot of sugarcoating when it comes to AI being useful one day.

That's not bad at all!

PHP and a plain-text protocol aren't what I associate with high-performance, but I guess it's relative.

100 %. Firewalls basically do nothing. If you are running vulnerable software, it won't help. If not, it's not helping, either. It basically only helps in the rare case that you have spectacularly misconfigured something. On the other hand, if a firewall is blocking automatic software updates, it's actually dramatically lessening security.

Judging by his recent podcast appearances, he is probably still fine. But it's also clear that he won't be for long, and it would have almost felt cruel had he been serving as president in his final years. He seems like the type of guy who'd rather spend time with his family and grandchildren.

You're constantly missing the point. Artistic merit has nothing to do with your feelings about a particular work or how it was made. That is your personal preference or opinion, whatever. People decide on artistic merit organically as a collective.

The lost artwork was discovered almost exactly ten years ago, in 2014. Are there any new developments on this front? If the article contained anything new, I missed it.

FOMO, probably. Just like any other tech bubble, really.

I thought I said it. Its main purpose is to eliminate guesswork.

Yes, it's misunderstood.

A lot of us were doing some form of agile before someone came up with the word agile.

I'll give you an example:

Imagine the process of creating a retro video game. First, you might write a function that clears the screen. Next, you'll get something simple drawn. Maybe just a simple block. Then you write a function that allows you to move the block with a controller. Then you swap the block with some sort of character animation. Next, you'll write a routine to draw girders. Then you'll implement gravity. Etc., etc. You'll just keep iterating until you have a well-balanced and engaging game.

That's agile. You may have had a lot of the game designed beforehand, or none. In fact, you could even create the game design documents with the same method. Make a quick first draft, get feedback, make adjustments, and repeat. And then keep going back to it whenever you notice something doesn't work or some better idea comes along during the course of development.

Originally, agile had nothing to do with rigid meeting schedules. Or "adapting to changes," or whichever buzzwords people like to throw around.

Which is funny because agile allows for almost anything. Even the Scrum Guide says that the process must be adjusted to the needs of the project. It's not even an option. Yet scrum meetings are treated as if they were religious gatherings.

> At least within my organisation, we do not design anything up front. We’re agile

Agile doesn't mean forgo design completely.

> We don’t think about proper api modelling. We’re agile.

See previous point.

> We also do barely any testing. We’re agile.

It's difficult, if not impossible, to be agile without testing. If you want to move fast, you want to be confident that your latest change didn't break anything.

> We do rewrite the UI a dozen times based on user feedback. After all, we’re agile.

Sounds like you built a complete UI before any users could give you feedback. That's the opposite of agile.

Agile means delivering very small increments to users at a frequent pace, allowing them to provide feedback early and often. When done correctly, this eliminates a significant amount of guesswork. However, if the client (or, indeed, your own team or management) does not fully understand the process, things can go spectacularly wrong in many different ways.

That's all kinds of wrong. A stand-up is supposed to take 15 minutes max, and if no one is stuck with anything, it shouldn't take more than just a couple of minutes. It's not supposed to be a "status meeting" at all.