@ken999 The cynical view is that people are trying to cash in on a trend, slapping AI into whatever they can and trying to make a buck.
The non-cynical view is that LLMs are the most transformational technology to emerge since cloud services, and they’re easy and cheap to use. An explosion of new ideas for AI-enabled applications combined with the low barrier to entry for individuals or small teams to build them means a huge number of new AI-enabled SaaS products.
Which view is dominant? I tend towards the non-cynical. I see most of the “AI-washing” happening with more established products where companies feel like they need to generate some buzz or risk falling behind. IME the truly “new” AI products I see are (by and large) authentic uses of new AI technologies built by excited developers who are earnestly trying to create something new, even if the net result is a lot of “ChatGPT wrapper” products.
@ken999 Exactly! More than 20 years after Internet development, softwares become more segmented and mature. AI brings new possibility of software development, allowing software developers seeing new opportunities.
Everything new brings a wave of followers, the early bird gets the worm
That being said, AI already has (and is continuing) to change the world.
There are a heap of existing solutions implementing great quality of life improvements using AI/ML; there're also a small number of companies building compelling products that aren't just content generators.
Like any bubble, most companies that don't have good economics, market fit or defensible offerings will die. Look at the "dot com" bubble. A few great companies will be potential unicorns, there'll be a decent amount of small companies that take advantage of AI/ML/LLMs, but most will die.
(I work for an existing SaaS that already is profitable, but is using LLMs and other models to improve our offering)
@ken999 The majority of "AI powered" apps can probably be made just as easily without the API calls to OpenAI but people want to cash in on the hype train. To me it feels super slimy, but hey morals are subjective so no hate if people want to get theirs.
I think we'll see this for another couple years, especially for consumer apps, until whatever the next trend is or if the AI cultists are right and it achieves "takeoff" or whatever then it won't ever stop.
@613jono Its not slimy if you are implementing something that someone else doesn't have the time or skills to do, such as building tools to wrap API calls.
It's not slimy for a Web developer to be paid to implement stripe, or to hook in an S3 data storage solution or whatever else. These are all just API wrappers to at the end lf the day.
People pay for convenience, that's how SaaS offerings work. The really simple or lazy wrapper offerings will probably fail to stand the test of time as OpenAi make their GUI tooling more powerful, and that will be justice served as far as your concerns go, I would think.
@watchdognp To clarify, the thing I find slimy isn't people using LLMs when it fits their use case and actually adds value. I personally don't understand the craze with the "chat with PDF" stuff, but obviously other people find it valuable. I can't think of a way to do that nearly as effectively without an LLM. Good example of an appropriate use case. Not slimy.
However, for other use cases like form builders or web scrapers, or todo apps or any classic done-to-death SaaS idea where the logic is pretty much straightforward and solved, slapping an openAI call on it to say "powered by AI" or "AI tool" to get in on the hype is what I see as slimy.
This is also why I couched all of that in saying morals are subjective. I'm not prescribing that viewpoint to anyone else, just that I personally would not feel comfortable doing it, and that's my "problem" for missed opportunity I guess.
