I sure hope you're right though. To me x402 can subsidize the cost of producing and serving content with payments from the systems that are using it and giving the least back. It can mean more free content for humans with the right payment model.
That's only the consumer side of it though. As the post states:
> Should a developer[...] elect to register themself with Google as a “verified” developer, they should expect to sign up for an account and pay a fee, surrender detailed personal information and upload government-issued identification, and then proceed to register the identifiers and signing keys for all the apps they intend to distribute (now or ever).
Those are big impediments to open development. The agreement developers sign states:
> 6.5 If You violate any of the Terms or if You distribute malware or other harmful applications, Google may terminate Your access to the ADC…
But they don't actually define "malware" anywhere in the document. Search HN if you want to hear horror stories about how google handles loose definitions and peoples' accounts.
The big difference I see is in the chip. The PowerPC arguably had its benefits (vectorization) which made it super attractive for bioinformatics, etc., and a lot of that software was Linux-based. People could either buy a super-computer or a G4 (or a cluster of G4s) and get the work they needed done for a fraction of the cost. MacOS (and OSX) were behind on a lot of this stuff compared to Linux then.
Today from what I see the M3-M5 chips are a big leap forward compared to their competitors, and it just happened to hit at the same time LLMs became popular. I imagine there are some similar, specialized needs with the M[1-5] chips that might benefit from Linux but with OSX's stronger BSD underpinnings it's a different world.
They've been like this long before they were a digital services company. 25 years ago Yellow Dog Linux (another RPM based distro) had the same challenges working with PowerPC. The scientific community was clamoring for an open platform to use their native Linux software with the PowerPC, YDL filled that niche, and Apple watched their struggles supporting Linux on their platform with detached amusement.
Mistral is generally winning the fights it chooses to fight and that's what they need to do.
Instead of looking at what EU's economy could contribute towards a SotA model it's more accurate to look at what France's economy could contribute, then compare that to the US or China. The scale isn't there. Instead what I like to see is what they can accomplish with that lower scale, and it's stuff like Leanstral, Voxtral and other niche products.
We are not Mistral's target audience. For instance I don't know if Leanstral performs the best as a "formal proof engineering model optimised for automated theorem proving and autoformalization" because I don't even know wth that is or who else does it.
Mistral themselves focus more on b2b; financial services, manufacturing, stuff like that, and they get some big clients that way.
Despite not being their target, I started using them because they have many open models. I continue using them because, yeah EU, but also because the community is great and the tool makes me think more than Claude does. Last, I stick with them because they are one of the few AI companies that are up-front about their environmental impact and are actually trying to minimize it while still providing a decent product.
> How much difference is there between beacon fire and fiber optic cable?
Like ... a lot? Now if RFC 2549 had been around back then you could get the same point across without trying to describe how information, rather than nectar, might flow through the equivalent of a butterfly's proboscis that happens to stretch around the world.
I think the takeaway is that swarming behavior is not difficult. It's relatively simple, it's cheap, and it's going to grow in effectiveness and smaller actors are going to gain access to it. That's going to be a problem for war-makers who don't take it into account. I think the pilot expressed awe at it because it was part of what beat him.
Man I despise Facebook, no two ways about it. But one time our dog slipped her collar and escaped while we were out of state. We got back late in the evening after she had bolted. We were up all night searching for her and we posted to the Ft. Collins Lost & Found group. We got a few hits that night, and by the next morning there was so much participation that we could trace her movement through town with an accuracy rate of a few blocks. The latest picture was at the intersection of a side-road and the state highway as it left town, so I was down there, looking around and freaking out that she was street pizza, when she booped me on the back of the leg. Don't know how I missed her but I sure am glad she found me.
When social media works right it's an amazing thing. It's just too bad that we don't use 90% of it right. As an example I've never given back to that commons when I see a stray pet.
It's probably too annoying for most, but I had a similar problem with KoReader and library books because those too depend on DRM. I ended up keeping KoReader as an app, then just not loading the app when I was reading a library book. It wasn't too bad.
KoReader also gave me a lot of freedom to manage the bad battery life of my Kobo Sage before it died of other causes. Definitely worth the extra cognitive load of dealing with the two experiences.
Is this a yes-and kind of thing or a no-but? I'm asking because maybe water serves as a good indicator for the other resource consumption issues around data-centers. Most of the water issues are are directly related to the energy and rare-earth issues you're talking about. This paper covers the scope of energy, consumption, and downstream water usage for servers:
meet.hn/city/us-Longmont