"We're often asked: 'Will AI-generated images replace photographers?' The answer is yes... for certain kinds of photography. For example, stock photography might be replaced by AI-generated images in the next 3-5 years." email from Eric Yang, CEO of Topaz Labs.
Good lord let's hope not. The implication of these kinds of statements about AI is that everything that could be done has been done. That's because the current AI implementations all work from knowledge of things that they've scanned. AI is not really a creative thinker. Even those so-called "hallucinations" you might hear about are based upon what was scanned, not completely made up from scratch.
One reason why stock photography exists is because if everyone uses the same images for their landing pages (or whatever) on the Web, then all marketing starts to look alike. Thus, if you subscribe to any stock photography market (disclosure: I do, though I rarely use such images, and then mostly only as ornament) you discover that the stock goes through trends, almost like fashion. I don't think AI is going to be dictating fashion. Indeed, to a large degree, one reason why fashion changes is that people want to avoid what everyone else is doing.
But this does suggest that if you want to stand out as a photographer you need to approach photography from your viewpoint and create new and compelling works, not simply copy others. Funny thing, that. I've been preaching that for decades now.
"I think that many in the smartphone generation are happy with one or two lenses...we don’t see the same numbers of people shelling out for four lenses as we used to..." interview with Fujifilm UK manager Theo Georgiades in Amateur Photographer
I'd describe this differently. We go through cycles in terms of camera buying, even as we go through technology cycles with cameras (e.g. film SLR, DSLR, mirrorless). Early in the new technology cycles, many of the first to move are those who are already using cameras but see an advantage to the new platform. That certainly happened with DSLR, and I believe it happened with mirrorless, as well. Current camera users saw something in the new technology that made their photography life better and jumped to the new form. It isn't that consumers discover the new form and then the established customers slap their wrist to their forehead and go "doh!"
After awhile, the excitement from the prosumer/pro users already having made a shift tends to get the attention of consumers, and you eventually get a new crowd buying, and buying mostly in the lower half of the lineup. That happened in the 70's with film SLRs, the 00's with DSLRs, and is now happening in the 20's with mirrorless. These folk don't have lenses, and don't yet know how lenses can be used to distinguish their work. So obviously they just buy the kit and use one lens for awhile.
It's actually up to the marketing departments of the camera companies to somehow convince those new users that they should be considering getting other lenses. Hmm. Who's doing that level of marketing at the moment? Bueller? Bueller? (Bonus aside: the lecture that Ben Stein, the economics teacher who say those famous words, was about tariffs and what happens when you apply them broadly, as with Smoot-Howley or...)
If we talk specifically about the smartphone generation moving to dedicated cameras—and indeed that's happening, just not in the huge volumes Tokyo would like—they're actually used to one "lens" being built into their device. Even the three-lens setup on the current iPhone Max Pro is essentially operated most of the time much like a mid-range zoom by the people I've watched.
That actually explains why cameras like the Fujifilm X100VI are so popular: it's just a "better" all-in-one solution than a smartphone (you "zoom" with its 40mp sensor by invoking real-time crops). Ironically, most of the X100VI users I've watched don't use that really sophisticated viewfinder. We'll see plenty more camera companies targeting this kind of device without really understanding why it's popular and then missing a key point or two in their design. I'd rather see someone such as Nikon make a Z9-generation Z30II but also figure out how the lens angle should be played with it other than "buy the kit." (One hint: you can get more telephoto from pinch to crop (or button to crop), as 8mp is more than enough for the uses these cameras are put to; but there's much more you need to do than just that.)
So as to that second part of the quote, the reason you don't see new-to-dedicated-camera users buy more lenses is because you, the camera company, have failed to show them how that becomes a useful extension to what they're already doing. Marketing is failing the camera companies big time these days. This is a marketing fail, not a completely new type of user that doesn't want lenses.
Finally, there's this: the overall market is smaller than it was. If there were one million users buying four or more lenses 15 years ago, there are only 300,000 or so doing so today.
"The problem with such reviews by individuals is that they are far too subjective. They only tell you about that one person's experience.” —dpreview forum
The implication of the first sentence is that reviews should only be done by organizations, and that organizations aren’t subjective. My experience, however, is that the larger the organization doing reviews, the more likely that their judgment is contaminated by other factors, including but not limited to how the group is funded. In the camera industry, we have a long history of magazines and others pulling punches on reviews so as not to offend advertisers.
Reviews contain basically two components: facts, and interpretation of said facts. For instance, someone could write “the camera produces 24mp images” or they could write “the camera produces 24mp images, which is (far) below today’s standard.” The first clause is a fact, the second is an opinion and based upon a second opinion (what today’s “standard” is).
You’ll note that I split my reviews into a few parts. The “What is it?” portion tends to stick to factual information about the product. A knowledgeable reader can form their own conclusions from that information (e.g. maybe they think f/6.3 is too slow an aperture for their work). The “How’s it Handle?” Is 100% my own subjective experience, based upon now 50+ years of using cameras from all makers, and a great deal of experience working as a professional. The “How’s it Perform?” section is a combination of facts and opinions. A buffer size can be reported in factual numbers, for instance, but subjectively that number works differently for different cameras. I’ll give an example: if a Zf-type camera had a 20-image buffer that might be fine for it while if a Z9-type camera had that same 20-image buffer it would be highly limiting. In order to understand that difference, you need a competent, reliable, and consistent reviewer who writes clearly.
Which brings us to the real thing about whether to trust a review or not: is the reviewer competent, reliable, and consistent? If you’re just browsing around the Internet and hit a review from someone you don't know you can’t make that assessment; I’d say that you have to take what they write with a grain of salt until such time as you can determine that they are competent, reliable, and consistent.
Now let’s deal with the second sentence of the quote, in particular the “one person’s experience.” The implication here is that someone else would have a different experience with the thing being reviewed. That’s possible, for sure, but the way that happens most often is that the other person’s experience biases their experience. Let me explain: if you hand a Canon camera to someone who only uses Sony cameras, their “experience” with the Canon is likely to be poor from the start. Controls are in different places, names are different, menus and customizations are different, and there’s a lot to learn about the Canon technology before you can master it (the implication is that they already learned the Sony technology and mastered it). Meanwhile, a reviewer who’s had broad use of Canon, Nikon, and Sony products actually has a better ability to describe where the product in question succeeds and fails. Yes, it’s an opinion based upon experience, but it’s an informed opinion.
More often than not, what I’ve found is that when people are objecting to reviews, they’re objecting more on the personality of the reviewer (including me) than the review itself. You see this all the time on the Internet fora. Indeed, it appears that the thread that generated this quote was removed by a moderator, likely because the discussion got into making strong negative comments about the reviewer, not the review.
“Company X should release a Y variant and let the market decide.” —common post, most recent of which was on dpreview
That’s not typically how successful businesses work. Throw spaghetti on the wall and see what sticks is a tactic that has been used by some, but it’s also an indication that they don’t know what business they’re in or who their customers are. With cameras and lenses there’s usually a two or three year creation process from idea to entering manufacturing, and R&D is one of the Japanese companies’ biggest expenses, particularly with lenses.
Nikon Imaging, for instance, is projecting a 24.5b Yen R&D cost this year against a likely 47b yen gross profit. Put another way, their profit is about twice what they put into development. They can’t afford to throw spaghetti at a wall; that increases R&D costs without necessarily generating profit.
This gets back to a common theme I’ve been writing about for three decades now: do the Japanese companies properly understand their current customers as well as their potential new ones? I’d argue that they are far less efficient at that than they should be. They miss clear easy pickings that I can see through my own user surveys, and they do silly things like leave out one or two features from cameras that should be configured identically (there’s no financial benefit to leaving HEIF out of the Z9, for example, or Cycle AF-area out of the Zf).
The most recent occurrence of the above quote had to do with wanting Nikon to put out a 400mm f/4.5 or 500mm f/5.6 with a built-in TC. The interesting thing here is that I tend to agree. If Nikon were to bring the 500mm f/5.6E PF VR over to the Z-mount, they certainly have the “air space” to stick in a TC and make it different. Moreover, it’s really clear to me that there’s a huge advantage to providing built-in teleconverters on the more expensive, exotic lenses, particularly for the 45mp users, as that gives you prime, prime+1.4x via teleconverter, prime+1.4x+1.5x via DX crop, which is effectively as good as a zoom much of the time. Even better, putting out a 500mm f/5.6 TC PF VR S might be enough to get people already using the F-mount version on the FTZ adapter to bite on the new version, too, as it gets rid of two extra mounts (you have to use an F-mount external TC and an FTZ adapter to get the extra reach now) and simplifies the switchover from one focal length to another.
But this isn’t “let the market decide.” It’s “the market would embrace this.” While that seems a subtle difference in wording, it’s the difference between doing something without knowing what your user might want versus doing it because you know what your user might want. Successful businesses always opt for the latter.
Finally, there’s this: some keep claiming that there’s no benefit of an internal teleconverter to an external one. Most of the time that argument centers on optical difference, not practical difference. Having now used the 400mm f/2.8 TC VR S for several years, I can tell you that being able to photograph at 400mm, 540mm, or 810mm with just a flip of a switch (TC) or a press of a button (DX crop) is vastly different than unmounting the lens, digging the teleconverter out of the bag, mounting the teleconverter, and then remounting the lens. It also doesn’t introduce a weak link in the lens/camera connection and another thing I have to carry, track, and clean. Personally, I’d pay for an internal teleconverter option on any long telephoto Nikon made.
This is another of those “learned things” that turn out to be wrong (“external teleconverter is more flexible”). Just because something has been done one way in the past doesn’t mean there isn’t a better way of doing it in the future. And that gets me back to my point: the camera company has to understand its customers well enough to know that something like a built-in teleconverter is highly welcome, not just throw the option out there and see if they like it. Look carefully at the most popular long telephoto lenses of recent times and you’ll find all of the internal teleconverter lenses in that list. The customers have already spoken.
_______________________
Bonus: Let's go back to that review quote. So why wouldn’t we trust an AI response to the question of “which camera is best?” Simple: AI is not overly competent at the moment (depends upon how well it was trained), not reliable (sometimes makes up answers, called hallucinations), and is not consistent (same question asked by different people at the same time get different answers).
Search isn’t any better, mostly because it’s no longer reliable (the search prioritizes how the search engine will make money, not your answer).
None of this is new. Throughout the history of media you find the same thing: the outlets that spend the time, money, and effort to be competent, reliable, and consistent are the ones that people eventually find and stick with. Indeed, the long-term viability of media tends to center around this notion.
