Linearity used to be a fairly straightforward iPad design app. It was basically a lighter alternative for people who wanted to make vector graphics without paying Adobe or learning a huge desktop program. Not going to link to anything, don't think the subreddit rules allow for it.
but like EVERYONE else it has suddenly reinvented itself around AI.
Maybe the product is useful. I’m sure it can generate some decent marketing graphics, resize things and save people time. Claude Design feels a 1000% better. But the whole thing feels less like a company developing something meaningful in AI and more like a design app realising that “AI” is where the enterprise money is.
Linearity does not have its own LLM. It is taking models and technology built elsewhere, putting them inside its existing design software and presenting the result as a new AI platform. There is nothing automatically wrong with that. Almost every AI startup depends on someone else’s model.
The annoying part is the gap between what these companies are actually building and how they talk about it.
A design tool adds a prompt box, connects to outside models and suddenly it is talking about changing how creativity works. Everything becomes an “AI engine.” Templates become intelligence. Brand guidelines become an intelligent brand. Automation that would previously have been sold as a useful feature is now treated as an entirely new category of technology. At some point we need to ask what exactly the company has contributed. Or?
Claude Design is much more interesting to me because it comes from the opposite direction. Claude is already a general model that can reason across writing, research, code, documents and design. The design part has the potential to become one part of a much broader working environment.
That seems like a more believable future than paying for dozens of separate AI wrappers. One for making banners, another for presentations, another for logos, another for social posts and another for resizing the same social posts.
This also connects to the larger problem with AI right now. We are creating an economy where a handful of companies train the models and thousands of smaller companies sell access to them through different interfaces. Each one adds a monthly subscription, a credit system and a layer of marketing language claiming that it has transformed an industry. Most of them have not transformed anything. They have made one existing task slightly faster.
Again, that can still be valuable. I would happily use a tool that turns one design into ten correctly sized versions. But saving twenty minutes is not the same thing as reinventing creative work.
There is also something bleak about the obsession with producing more content. Companies already publish far too much material that nobody wants to read or look at. AI is being sold as a way to produce even more of it, faster and with fewer people.
The bottleneck was never just the designer taking too long to make the banner. It was usually that the campaign was uninteresting, the message was vague, nobody had made a clear decision and six people needed to approve it.
This is why I find Claude Design more promising, even though it will obviously have plenty of problems of its own. The interesting possibility is not simply that it can generate an image. It is that the same system could understand the research, the brief, the product, the copy, the design and perhaps the eventual implementation.
Linearity and others feel more like an existing software company attaching itself to that change because the old category of “nice iPad design app” was not going to produce the same valuation or enterprise pricing.
[link] [comments]