hbosch

>I'm sure the same is true of doctors too.

And police officers, teachers, race car drivers, and engineers too. There is an implicit assumption that most professionals all meet some unified bar of competency but that is far too often not the case. Like most anyone else, dentists and doctors are out to get as much money as they can for as little work as possible, and that is the system we currently endure.

When I find someone who gladly tells me I don't stuff to pay for, I keep their number. On the other hand, it's getting more and more difficult to find an AC repair guy who isn't trying to sell me a new unit every maintenance visit.

We are talking about Microsoft employees.

Naturally, there’s an old This American Life episode about this which focuses on young international people who work summer at a highway rest stop.

Ep. 388

Every decision a human makes has some level of intent. AI cannot intend, it can only suppose.

How much will a license cost? Can it be earned? If granted, granted by whom? Can I use it across models? Does it need to be reassessed regularly? How often?

If any of the above are more convenient than paying an engineer you aren’t any safer, the enterprise paying is just putting money in someone else’s pocket.

>There are enough people who would pay any money to have an electric Ferrari. The fact that it's a rather everyday car—and not a supercar—makes it a very attractive option for rich people who need to show off.

In case it wasn't clear, the Luce is a 1,000+ HP car and will cost over $300,000 USD.

Unity enabled a flood of slop games long ago. Dreamweaver enabled countless slop websites. Photoshop delivered us heaps of slop images. Amazon delivers thousands of slop products from slop manufacturers every day.

The slop isn't coming, it arrived decades ago. The Pandora's box of slop is already open. Maybe AI widens the aperture, but if you cannot handle the discernment required to separate slop from something useful or meaningful, that is your problem.

If I download Blender today, as a true beginner, is what I make extraordinary? If it's not, does that mean I am not allowed to use Blender? What if I want to use Blender and I am not interested in making anything extraordinary? What if I want to use Blender to make a stupid little iPhone game that no one will ever play? Is that considered extraordinary, or not? What is this criteria?

The truth is, the vast majority of art is not extraordinary, whether it comes from a canvas, a typewriter, Photoshop, or Blender. That is as true for AI as it is for humans. Likewise, the vast majority of people who kick a soccer ball will never be extraordinary soccer players.

I firmly believe that tools which enable people to get closer to their goals are always a good thing. The concept of what makes something "extraordinary" does not come from the maker, or the tool, but from the beholder. It is the audience's job to discern what is and isn't "extraordinary", not the makers'.

There is no scenario where more people using Blender is bad for Blender.

Homogenous doesn't work when you're Google and your product comes out looking like Microsoft.

The solution is an open, flexible, scriptable and drawable canvas where design and code co-exist in exact harmony. Design changes directly modify front-end code, and front-end code directly modifies design equivalently. I see the endgame as a model where designers and FEE's are co-owners and co-authors of the front-end with zero handoff.

I used Claude Design to see how it'd spit out a design I already had been working on for some weeks, given a dense enough prompt and a decent requirements document (I did not feed it visuals). I thought the output was pretty good! It didn't match the style we're after at all but it did do some logical content grouping and made some IA decisions I decided to pull into my own explorations. Overall I left with a good impression.

And then I was scrolling Twitter, and saw someone else post their own "success story" and the design was nearly identical to the mock up Claude Design made for me. Lol. The homogenization problem will continue to plague tools like these to some degree, much in the same way AI generated text or code or imagery has a sort of homogenous tone or feel to it.

Fonts are software. You can program them such that any two letters beside another can render uniquely. This is most common with ligatures like (e.g. fi -> fi) but also, say, swapping a colon from baseline oriented to centered if between 2 numbers, and so on.

>Has no one tried this before?

This is a great execution of a very common font practice.

When better standalone LLMs got "web crawling skills" integrated, it pretty much destroyed the need to ever lean on PPLX again. Perplexity is actually not a bad product, but other services like ChatGPT and Claude can do it's best thing pretty good, and do other things much better.

One thing I noticed is that whatever harness PPLX wraps around the models, the output is noticeably lower quality in aggregate. I assume some kind of token compression being used before passing your query to a given model but to my knowledge that's never been proven or confirmed?

Anyways, I get the most value out of coding and PPLX has seemingly pivoted away from that. Probably a good play to not try and compete directly with Claude Code/Codex and find a better niche, but I am not sure who or what their market is. Lovely design, however.

Being Microsoft, you'd think they would just offer a public Teams server instead? Not that you'd get more traction with it, but at least it's in-house and theoretically they would be motivated to build integrations on top.

Experience hardens crystallized intelligence.

"Fluid intelligence" is not very valuable when it comes to long-term decision making.

>And if you can afford business class [...] The meals are restaurant quality and the full recline?! I hardly want to disembark!

Let's settle down. This kind of biz class experience is almost certainly unique to international travel. Flying "business class" from ATL to SFO might get you a plate of microwave slop and an extra 15deg of incline on almost all domestic jets. Once in a blue moon you'll get a modern plane with the diagonal seats. One less person in the row, though.

Paying for business class domestically is almost always a sham by my experience.

This is a bit ridiculous in practice. The reality is that products have many, many vectors of experience. Like a house does. If you have a broken window and a leaky pipe, you can hire 2 different people to fix both of those things separately...

>This will not end well for Disney

I have faith that the Parks Imagineers will soon be installing Sora Stalls in and around every attraction in Disney World.

So, the centrifugal force of head-banging is why metalheads grow long hair?

"Wanderful" would be a better name.

Are you sure? While Amazon doesn't own a "true" frontier model they have their own foundation model called Nova.

I assume if Amazon was using Claude's latest models to power it's AI tools, such as Alexa+ or Rufus, they would be much better than they currently are. I assume if their consumer facing AI is using Claude at all it would be a Sonnet or Haiku model from 1+ versions back simply due to cost.

It may have happened. There are already many users saying their "created in" locations were incorrect. Thus the rest of my comment: trust is binary. We can either be 100% certain the data is correct, or we must assume it is never correct.

The same head of product quoted in the sib comment admits that "for a small set of accounts the location data was incorrect". Given what we know about Twitter's relationship with the government and this administration in particular, you're simply left to do with that information what you will.

I personally do not trust Twitter, or the government, very much. I also would not be surprised if some government accounts were created at various embassies around the world or through strategic VPN networks, or if general business is conducted through a darknet-like node system which includes allied endpoints. To me those are more plausible.

Does this make the fact that it showed as Israel, was disabled, and is now "corrected", mean that this feature is good or bad?

Either we can trust all location data all the time, or we can trust none of it. We cannot expect Nikita Bier to swoop in on every suspicious tweet and try to educate us on IP range changes and DNS glitches or whatever.

Furthermore is it more likely that a small set of special accounts seemingly never collected location data on signup, or that for a small number of accounts X simply modified that data post-hoc?

It wouldn't be difficult to include a little "" icon on each post that, when hovered, shows probable location or the creation origin. It's true that VPN's will often muddy this info, and even things like phone number confirmation for 2FA or billing address for Verified badges can't be 100% reliable... but it shouldn't be too difficult to say in those cases "User likely masking location" if the pattern is detected.

>It’s widely known and reported this is happening.

I am not sure it's "widely known and reported" that the official Twitter account for the Department of Homeland Security was created in, and is run from, Israel.

I used it to mean that it's inarguably true and unrelated to opinion.

The page is objectively and substantially better in every aspect of legibility with CSS loaded.