It really depends on how much you reduce costs. If you reduce costs enough, you can have increasing supply even in the face of falling prices. This argument sounds like one made by a hedge fund protecting its real estate investments.
The reality is that the housing market in the US (and most countries) is heavily distorted by government NIMBY regulations. Because of this, it's reasonable to expect that there is actually a lot of room to reduce the $/sqft if the market can build housing in general. Current costs are inflated by being forced into specifically prescribed solutions designed to grow the wealth of developers and landowners.
Different states are very different in this regard; Austin Texas is considered the NIMBY-est city in Texas, and it's still building lots of houses even as prices drop, and have been for a couple years now. Other states do (mal)function as you describe, but not all of them.
Price reduction against builder margins would be a more reasonable way to look at it than raw price change, and additionally factoring in the degree of financing vs cash and the cost of financing would possibly shed more light.
Stable firms, not out over their skis, can afford to entice buyers while still making a profit. Just like in a startup, the cost of money and the degree of leverage sets the [minimum] speed limit on the runway.
Machine translation is a great example. It's also where I expect AI coding assistants to land. A useful tool, but not some magical thing that is going to completely replace actual professionals. We're at least one more drastic change away from that, and there's no guarantee anyone will find it any time soon. So there's not much sense in worrying about it.
A very similar story has been happening in radiology for the past decade or so. Tech folks think that small scale examples of super accurate AIs mean that radiologists will no longer be needed, but in practice the demand for imaging has grown while people have been scared to join the field. The efficiencies from AI haven't been enough to bridge the gap, resulting in a radiologist _shortage_.
> Machine translation is a great example. It's also where I expect AI coding assistants to land. A useful tool, but not some magical thing that is going to completely replace actual professionals.
I can say from experience that machine translation is light years ahead of 15 years ago. When I started studying Japanese 15 years ago, Google Translate (and any other free translator) was absolutely awful. It was so bad, that it struggled to translate basic sentences into reasonable native-level Japanese. Fast forward to today, it is stunning how good is Google Translate. From time to time, I even learn about newspeak (slang) from Google Translate. If I am writing a letter, I regularly use it to "fine-tune" my Japanese. To be clear: My Japanese is far from native-level, but I can put full, complex sentences into Google Translate (I recommend to view "both directions"), and I get a reasonable, native-sounding translation. I have tested the outputs with multiple native speakers and they agree: "It is imperfect, but excellent; the meaning is clear."
In the last few years, using only my primitive knowledge of Japanese (and Chinese -- which helps a lot with reading/writing), I have been able to fill out complex legal and tax documents using my knowledge of Japanese and the help of Google Translate. When I walk into a gov't office as the only non-Asian person, I still get a double take, but then they review my slightly-less-than-perfect submission, and proceed without issue. (Hat tip to all of the Japanese civil servants who have diligently served me over the years.)
Hot take: Except for contracts and other legal documents, "actual professionals" (translators) is a dead career at this point.
Also, Google Translate is really not a particularly good translator. It has the most public knowledge, but as far as translators go it's pretty poor.
DeepL is a step up, and modern LLMs are even better. There's some data here[0], if you're curious - DeepL is beaten by 24B models, and dramatically beaten by Sonnet / Opus / https://nuenki.app/translator .
Just a note on the radiologist part, the current SOTA radiology AI is still tiny parameter CNN's from the mid-late 2010's running locally. NYT ran an article a few weeks about this, and the entire article uses the phrase "A.I.", which people assume means ChatGPT, but really can refer to anything in the last 60 years of A.I. research. Manually digging revealed it was an old architecture.
We don't know yet how a modern transformer trained on radiology would perform, but it almost certainly would be dramatically better.
> We don't know yet how a modern transformer trained on radiology would perform, but it almost certainly would be dramatically better.
Why? Is there something about radiology that makes the transformer architecture appropriate?
My understanding has been that transformers are great for sequences of tokens, but from what little I know of radiology sequence-of-tokens seems unlikely to be a useful representation of the data.
On the surface, radiology seems like an image classification problem. That's indeed something a small NN can do, already 15 years ago. But it's probably not really all it is.
I can only imagine what people picture when they think about AI and radiology, but I can certainly imagine that it goes beyond that. What if you can feed a transformer with literature on how the body works and diseases, so that it can _analyse_ (not just _classify_) scans, applying some degree of, let's call it, creativity?
That second thing, if technically feasible, confabulations and all, has the potential to replace radiologists, maybe, if you're optimistic. Simple image classification, probably not, but it sounds like a great help to make sure they don't miss anything, can prioritise what to look at, and stuff like that.
> What if you can feed a transformer with literature on how the body works and diseases, so that it can _analyse_ (not just _classify_) scans, applying some degree of, let's call it, creativity?
Then it would make up plausible-sounding nonsense, just like it does in all other applications, but it would be particularly dangerous in this one.
That wouldnt be that much different than current CNN/labeling methods used on medical imaging. Last time I got a ct scan the paperwork had the workstation specs and the models/nueral network techniques used.
Having worked in medical software for a long time now, it's not surprising.
Medical software is purchased by upper management based on features. Usability is rarely considered, and hard to evaluate anyway since it depends on integration.
Hospitals themselves create this incentive because they will rarely if ever change their processes to fit more smoothly into a different workflow. The core purpose of an EMR is to integrate different parts of the hospital, though, so the only possible solution is a huge do-everything piece of software like Epic.
I spent years on a project to try to standardize exams between a major research hospital and its satellites. The ended up quitting about half-way through because they realized there just wasn't going to be enough political power at the main campus to convince the doctors at the satellites to follow their lead.
You'll even see this at smaller regional hospitals. We'll bring software and workflows developed in tandem with the top hospitals in the world, and they'll tell us that they don't care and they want to keep doing things the way they've always done them.
Everyone may hate EMR, but hospitals seem to be getting exactly the EMR they're asking for.
Obviously being LLM generated is a good data point because it shows that the OP isn't arguing against the statements of the review itself.
It's also good for the editor to know about. LLMs represent a new acute threat to review quality that they may currently be underestimating. I've literally heard of people bragging about using ChatGPT instead of doing reviews themselves. People who aren't LLM experts don't necessarily understand their limitations or that using them in this way should be unacceptable. The editors should know so they can improve the communication of review expectations.
The demand for human radiologists never actually went down, though[1]. There is improved efficiency due to technology, but it hasn't kept up with increased demand for scanning and increased patient volumes. There's also induced demand caused by efficiency increasing. Diagnostic AI tends to make different kinds of errors than a human, so even if it were better, both together is even stronger.
The "Radiologists are being replaced by technology" story has been repeated so many times by uninformed software developers that it has become popular wisdom, but the reality so far is that we need more radiologists than ever.
Ironically, radiology might be a decent proxy for what could happen with software engineering, but in the opposite way you intend.
Amusingly a poster above indicated they had issues with multiple Subarus, and another had issues with Hondas (which is what has worked well for me).
People on this forum want it to be a technology issue, and I'm sure it is to some extent, but I think the differences between driving styles is also a large component. If you're a careful driver that usually leaves a lot of separation, many AEB implementations will rarely trigger and you'll usually understand why when it does. Whereas if you're an aggressive driver that cuts between vehicles, leaves <1 s separation, etc., you will notice a lot of "unexpected" braking from these systems. Like the poster above that was upset the system slowed them down when they were nearly clipping a car exiting the road.
A lot of good comments here, here's one point I haven't seen: If you can, just respond tomorrow. Use the evening to exercise, visit friends, play trivia, etc. I find it can help you force your brain to see the comments as just one part of your whole life.
The creators of these images assigned the rights to adobe, including allowing Adobe to develop future products using the images. So yes, this is perfectly fair.
It's completely different than many (most?) other companies, which are training on data they don't have the right to re-distribute.
I think you are making a jump here. I’m not a lawyer, but your first sentence seems to be about why it is legal. And then you conclude that that is why it is also fair. I’m with you on the first one, but not sure on the second.
The creators uploaded their images so adobe can sell licences for them and they get a share of the licence fees. Just a year ago if you asked almost any people what “using the images to develop new products and services” mean they would have told you something like these examples: Adobe can use the images in internal mockups if they are developing a new ipad app to sell the licences, or perhaps a new website where you can order a t-shirt print of them.
The real test of fairness I think is to imagine what would have happened if Adobe ring the doorbell of any of the creators and asked them if they can use their images to copy their unique style to generate new images. Probably most creators would have agreed on a price. Maybe a few thousand dollars? Maybe a few million? Do you think many would have agreed to do it for zero dollars? If not, then how could that be fair?
No that isn't how it works with Adobe or any of the other big stock photo companies. The photographers or creators of the images still own the copyright. Both with rights managed and royalty free they aren't assigning rights to anyone else.
If you're an experienced individual contributor, the shift to remote work was great. You got less oversight at work and more free time in your personal life. This category is a lot of the hacker news audience, so remote work is popular here.
It was less good if you were a new contributor. You might find yourself with less development and struggling to break into existing cliques within the organization. You may not have built good practices in personal time management. None of this is malicious, it is just failures that are easier to have happen when working remote and not being careful.
Managers have an even harder time. Good managers work by building strong relationships with their team members, not with carrots or sticks. That's harder in a remote environment, due to the default-private nature of most remote communication. The lack of relationships hurts individual contributors too, who become even more like cogs in a machine.
Overall, I think the ability to retain experienced employees and hire from a wide range of locations outweighs these costs of remote work, but they are real and significant. For Big Tech, where acquiring talent is relatively easier due to salaries and name recognition, you can see why they might prioritize in-office work.
It really depends on how much you reduce costs. If you reduce costs enough, you can have increasing supply even in the face of falling prices. This argument sounds like one made by a hedge fund protecting its real estate investments.
The reality is that the housing market in the US (and most countries) is heavily distorted by government NIMBY regulations. Because of this, it's reasonable to expect that there is actually a lot of room to reduce the $/sqft if the market can build housing in general. Current costs are inflated by being forced into specifically prescribed solutions designed to grow the wealth of developers and landowners.