I think this could be something where the middle ground is the best option, this being, just make the rule that any listing with a picture that includes GenAI must also include the original un-AI’d photo right before it. This allows the lister to present the place as it is (important to the renter) and how it could be (important to the lister). I don’t think everything needs to be black and white
The pricing on this (it’s… fine?) might say more about the costs of running any frontier/borderline frontier model than any other insight. Either Kimi knows that the cost to run the premier models are being highly subsidized and will right course soon, or the cost for any borderline frontier model at this point is just flattening.
I have a young daughter and when I think about the most critical skills I want her to develop throughout her adolescence, communication is one of the most important ones that will prove to be valuable throughout her life. In whichever career she chooses, with friendships, personal mental health, partner relationships, etc.
She is smart, she is talented and incredibly curious and those things I really do not worry about. What will set her apart from the majority of her peers throughout life will be her ability to effectively communicate and interact with others in a way that is meaningful. It’s benefits go far beyond what most of us appreciate
However succeeding is increasingly harder and harder.
You can be the best communicator in the world but when the world only want to push their story to manipulate you towards their will then the only skill that counts is obedience or being the most aggressive manipulator. I wish that meaningful communication was a key to good and content life.
You still can have your bubbles in the universe though, there it is invaluable, of course. Isolationism on the personal level might be necessary, unfortunately.
Does AI end up being much closer to a Crypto style result late 2010’s-early 2020’s) or is it closer to the Web style late 90s landscape?
Both had somewhat similar early outcomes for a lot of the initial front runners with some experiencing legal issues and others collapsing due to sketchy finances. Crypto ended up not being as earth shattering as most believers wanted it to be but there are still many billionaires and large companies that exist and continue to churn profit. Will a similar fate exist for AI companies? A few large winners, lots of large/medium/small losers (with some taking the legal and financial hits) but a decently large industry (not earth shattering) survives and establishes to simply be a small piece of the greater economy rather than the primary driver as most AI truthers believe?
LLMs are actually useful, people are willing to pay for access to them, and they do genuinely enable things that were unrealistic or impossible before. The advances in image, video, and sound models since 2020 are also striking but likely won't be as transformative as LLMs.
That being said, I don't think it's unlikely that we'll see a plateauing of progress followed by a strong crash/correction in the market a-la dot com. The Allbirds situation absolutely has echoes of pets.com.
I also feel that commodification is coming for models, training/inference hardware, and software (e.g. CUDA), as it has for nearly everything else useful in tech. So I expect valuations driven by unique advantages here to be eroded over time (Think Sun and SCO after Linux on cheap x86 servers became the norm).
I think it will be like the 90's collapse but bigger. It must be noted that the financial markets have been awash in cash for some time since companies have been making huge profits since the 90's. So the money that will be lost will mostly be that of rich investors and large companies who were sitting on a bunch of capital they didn't know what to do with.
My Google usage has gone down fairly significantly in the past couple of years as I use whatever free models to do nearly all of what I used Google search for. The only thing that Google has me with right now is YouTube and Gmail.
For now, until they’ve Googled the product into garbage, AI is so much better at getting me exactly what I’m seeking. I wonder if when considering all the ad slop that comes with every Google search or website load, it might actually be more energy friendly, or at least comparable?
I really wish settings could be downloaded and implemented like transferring Bookmarks in browsers. That way I could take a setup of settings someone has created and simply mimic that by way of adopting all of the same settings in one fell swoop.
It would break the anxiety barrier of having to finagle with a bunch of different settings which exists for a lot of people believe it or not.Or Apple could just make it easier to implement these types of features much easier.
I imagine the ideal situation would be a hybrid. Trains for the bulk of transport and self driving cars for the large tail of odd trips. When you want to get home from your friends house at 1AM there is no traffic on the roads, and taking PT is a 1+ hour trip while uber is 10 minutes. There will never be super fast PT connecting every odd route that runs frequently all through the night.
Cars could be radically reimagined for this use case to be much smaller and more efficient to be more like an ebike with a box around it for weather proofing.
Why not? Public transport follows static routes, but there’s nothing to stop personalised routes based on whose on the public transport, especially while off peak.
Personalised routes is not really much different than uber pool. Maybe the government should run a public version of this. But I think there will always be some trips that are so low volume (particularly late at night) that it makes sense to just have a single occupant car rather than a whole bus moving one person.
reply