Nothing will happen, we already have kind of translation activities happen between "incompatible views" in real life, it works.

Besides, in weights there are no real incompatibility can happen, all is unified with same math chain through architecture. Different set of local minimums/maximums can have different strong/weak consequences but nothing incompatible.

Different consistent representations also mean that model probably can be simplier for the same task, sign of excessive capacity, imho

  >Besides, in weights there are no real incompatibility can happen, all is unified with same math chain through architecture. Different set of local minimums/maximums can have different strong/weak consequences but nothing incompatible.

How do you know you're not stuck in a local minimum, always re-generating the God (or worse, the Demon) configuration and thinking it's universal?

This argument that incompatibility is impossible has an easy counterexample by looking at loss of coherence between two networks. You can still mash the weights together (so they're not "incompatible" in the basic number crunching, sure), but the resulting neural net is meaningless.

By "incompatible" I meant not in a software sense, but like two logical axiom systems can be incompatible.

Very educative, thanks for sharing the story!

They are used a lot, and modern edit models kind of support base cases - depth, canny, openpose - as a reference image. I mean they understand "use image 2 as depth map", result indeed have same depth, etc. But interpretation sometimes arbitrary to model.

For the tasks where several control nets should be adhered (depth and canny and segmentation) dreaded SDXL is still unmatched. SDXL also the only one handling influence weight for each control net normally, this ability completely vanished on modern models

Control nets in early post-SDXL models like ZIT or Flux kind of work too, but still much worse in quality when combined. I suspect pure diffusion architecture of SDXL is really favorable to mixing control signals with precision, DITs probably can be trained to similar level for mixing signals - but it is much harder and never seen in practice. At least every post-SDXL model I tried is trained with single one in mind, judging from my attempts. This rant is from personal expirience.

Great explanation, thanks! Would be cool to see this for people BSky instead of ex-Twitter

To understand hate against Adobe one have to experience attempt to stop Adobe subscription from renewing. This is truly life-changing experience.

This article reads to me like that "leave the multibillion dollar company alone" meme

> Yet despite this not one of these vibe coded projects has been able to create a brush engine that matches Photoshop’s brush engine.

The article author is brush hurt over this. Oddly specific.

I would not brush them off like this though, maybe they are the dev working on photoshop's brush, or brush is just the most basic tool in such software that it makes an easy comparison.

The person is literally that meme. Quotes from their other posts:

> They are pretending to be angry at big companies because they think it will make them look heroic and noble.

> So when the angry mobs online tell us that we must irrationally hate a race, nation, party or corporation - well I am not joining that movement.

Poor poor Adobe, they only want to make you pay $20 a month, for an app which hasn't substantially changed in any useful way since the early 2000s...

If they take legal action and I hope they SCO themselves, along with the rest who treated the customer base with utter contempt, for at least 20 years.

This. I had to create a temporary card with my banking app (Monzo), then I could switch out the payment method on my adobe account. Then I could cancel the temp card and then finally Adobe can die whilst trying to charge my cancelled card. They wanted me to pay £80 to cancel, I’d had the account for 10 years so no idea what the cancellation related to.

For anyone in this situation - first upgrade your subscription to a different tier, wait a day for the account to reflect you are on a new test, then you cancel with the 7 day right to cancel. That way you avoid early cancellation fees

This feels like it is a brilliant loop hole, but knowing adobe there is a 20% chance that this backfires and you are now stuck with a higher tier of billing for a minimum of one more year.

This, and also as an amateur once-in-a-while user of 2 apps, I don't get why I have to pay the same price as professionals using 5 or more apps every day.

There's a lot of comments about how these projects could never be used seriously, but I don't use these programs seriously. I take photos for fun and I make posters for events. I'll happily take the vibecoded lightroom and photoshop for free and it runs on Linux natively.

For me it's the other way around. I tried darkroom and C1 but I found both of them a lot harder to use and less feature-rich compared to Lightroom. If I was a pro user using it every day I'm sure I could justify learning either of those programs (and learn them relatively quickly) but as a hobby user I wanna take photos, not learn photo library management and editing from scratch. The quality and quantity of free tutorials and presets for Lightroom on YouTube and other platforms is also much greater. Granted I haven't tried these vibecoded versions, but I barely trust the Adobe products not to corrupt my library, I probably wouldn't trust these either

They charged me $60 to stop paying them money.

Does the country you live in not have consumer protection laws?

Read this with Saul Goodman voice

I had to terminate the card… it was honestly easier, they have even been fined over their practices.

I never understood that problem. Why can't you cancel the credit card used for the subscription? (Yes, it's a bit of a pain to move other subscriptions to another card, but also a great opportunity to clear out some other leeches.)

> Why can't you cancel the credit card used for the subscription?

ianal - but I never understood that mindset. You're surely aware that a subscription is a contract, and terminating the payment method does not legally terminate it unless the company doesn't bother proceeding further?

Which may be the case for Adobe, but you're definitely in a legal gray zone in which I'm sure (again ianal) they can sue for damages if they want.

Except their contracts are literally illegal under several jurisdictions, which is why they don’t attempt to enforce them usually. Just because they wrote it, and you signed it, does not always make it legally binding…

I've told this story before but I had to tell their scum support that I had cancer before they'd cancel my account. They just flat out refused. It was almost impressive in a way.

Drone units of occupation forces of Russia are primary taegets for AFU and due limited control distance for FPVs, Russian drone units are serving at the zero line, actually. So they are sent to actual front lines constantly (despite what Russia officially claims)

Yes, drone operators yield bigger score than the tanks! Bigger score means better equipment for your unit.

May be Ethereum should fund 10000 agents to pursue new crypto math with provable strongness/hardness

I think much of the the doomer propaganda is paid for by by the half billion or so that Vitalik Butering gave to the Future of Life institute. It sounds like he wanted it spent on research more than advocacy and is now distancing himself from them.

Also Krita

Yes, why not. "If then ..." Is for "harness", but the answer to prepared question is for jev-like cheap and fast service

Worth to mention that with Claude and GPT this can be result of tournament sampling, which is part of text watermarking. Same answer for all Claude models kind of confirm it, imho.

So not something internal to model thinking.