It is a focus, data, and benchmarking problem. If someone comes up with good benchmarks, which means having a good dataset, and gets some publicility around, they can attract the frontier labs attention to focus training and optimization effort on making the models better for that benchmark. This is how most the capabilities we have today have become useful. Maybe there is some emergent initial detection of utility, but the refinement comes from labs beating others on the benchmarks. So we need a slideshow benchmark and I think we'd see rapid improvement. LLMs are actually ok at a building html decks, not great, but ok. Enough so that if we there was some good objective criteria to tune things toward I think the last-mile kinks would get worked out (formats, object/text overlaps). the raw content is mainly a function of the core intelligence of model, so that wouldn't be impacted (if you get get it to build a good bullet-point markdown of you presentation today it would be just a good as a prezo, but maybe not as visually compelling as you like. Also this might need to be an agentic benchmark to allow for both text and image creation and other considerations like data sourcing. Which is why everyone doing this ends up building their own mini framework.
A ton of the reinforcement type training work really just aligning the vague commands a user would give to the same capability a model would produce with a much more flushed out prompt.
We've reach a point of price stabilization and longevity for smartphones now that didn't exist for the first 10 year ramp. When every new model added fundamental capability, you always want to upgrade, with the sweet spot often being every other year. But now, with better build quality, batteries, and stabilization of features people will keep their phones for much longer. Or buy "new" models that are of older versions since the price/features have been acceptable to run most of the apps they care about for years now. Plenty of people still want the top end for similar reasons to why people buy design clothing, but we've reached a feature plateau. We hopefully are getting close to that with EVs. Seems like around 300 mile range standard was the key thing. Though improved AI driving could change that again.
The main issue with smartphones is software support, as it essentially acts like a built-in time bomb.
Buying an older-generation flagship model to get better features than a current-generation midrange model of the same market price isn't very attractive when it'll have to be replaced after 2 years instead of 5 years.
If we want to get nitpicky, Tesla no longer has the majority of US EV sales. Their market share has fallen to 42.3%. They still have a huge lead on the next closest brand however.
> If so, why aren't used dealers just including a battery swap in the price?
I think that is the main thing that needs to be figured out. I suspect the problem is that you need to get OEM battery replacements for older model cars and those aren't yet readily available or cheap. We are going to need aftermarket batteries to drive price competition in the market. The current car manufacturers aren't incentivised to support a secondary market when they are still focused on primary sales. Also not in the ICE market there is much more ability to scale capacity. The supply chain constraints for EVs, and batteries are much tighter, though that keeps getting better.
Battery swaps are never going to be a thing long term, even with Nio rolling it out in areas. It adds huge amounts of weight and complexity. You have to build electrical and coolant connectors which can handle large amounts of connects and disconnects, in areas that get mucky and interact with rain, salt, snow and ice. You have to build a chassis strong enough to take an impact but also support the additional weight and space that a removable battery takes up - think of how much bigger phones with removable batteries.
I have done 900 mile road trips in EVs with 150Kw charging (low by standards of newer EVs) and charging has been a complete non problem. In fact I have more problems with plugging my car in, going to the toilet and coming back finding that I've put more power into the car than I wanted.
Batteries are lasting 200k+ miles with 85-90% original capacity in so longevity is not a problem and charging is becoming a solved problems in an increasingly large portion of the world too.
You put this in the wrong place. "Battery swap" in this context should be read like "transmission swap". Hours of work replacing a permanent part. Nothing to do with detachable batteries.
Hybrids keep their value remarkably well. If each engine isn't spinning half the time they will obviously last longer. They could have a small enough battery that make hot swapping a lot more realistic.
That is the technicality here. Bullshit is getting spewed, but in most cases, direct falsehoods aren't gett reported. If you quote someone saying something untrue, the paper didn't present a falsehood, same with bias, omission, emphasis and misleading narratives or framings. If you avoid stating facts and just cite sources, you can maintain, that the media outlet didn't lie. But only in the limited technical sense of direct commission.
Transportation crime fear is compounded by another issue: "scary people." I've personally never witnessed a crime. But I've seen plenty of people that raised my hackles, usually they seem intoxicated or are exhibiting some kind behavior that may indicate mental illness. Are they going to get up and stab me? Probably not, but it sure seems like it could happen, and it sometimes (though rarely in terms of transite miles) does happen. I can intellectually dismiss other low prevalence issues in a way that it is hard to do with public transit, viscerally.
I know you meant "became wary" when you wrote "raised my hackles", but that phrase means "to (visibly) upset or arouse one's anger," which I'm sure is not what you meant. But it does speak to a large part of the problem: people becoming overly engaged with something that they should probably just acknowledge and be aware of, without changing their behavior significantly.
Crime hysteria seems like it gets people, who are unlikely to be victims of crimes but more likely to have outsize political influence, involved in law enforcement policy. Without being forced to dogfood the results of their own advocacy, you end up with policing rules written by people who rarely are forced to interact with police, and who are very scared of crime that never happens to them.
Well, that depends on what we are selling. Are you selling the service, black-box, to accomplish the outcome? Or are you selling a tool. If you sell a hammer you aren't liable as the manufacturer if the purchaser murders someone with it. You might be liable if when swinging back it falls apart and maims someone - due to the unexpected defect - but also only for a reasonable timeframe and under reasonable usage conditions.
I don't see how your analogy is relevant, even though I agree with it. If you sell hammers or rent them as a hammer providing service, there's no difference except likely the duration of liability
There difference isn't renting or selling a hammer. The difference is providing a hammer (rent/sell) VS providing a handyman that will use the hammer.
In the first case the manufacturer is only liable for defects, for normal use of the tool. So the manufacturer is NOT liable for misuse.
In the second case, the service provider IS liable for misuse of the tool. If they say, break down a whole wall for some odd reasons when making a repair, they would be liable.
In both cases there is a separation between user/manufacturer liability - but the question relevant to AI and SaaS is just that. Are you providing the tool, or delivering the service in question? In many cases, the fact the product provided is SaaS doesn't help - what you are getting is "tool as a service."
>It's hard to be positive about the idea of your skills getting devalued and getting kicked to the curb.
I think it depends where people build their own identities in the value stream. Do you see yourself as a product/hacker type person and writing code just is a blocker on delivering your vision? Building greenfield prototypes is now 100x easier! Do you see yourself as a craftsperson that brings years of experience to hard technical challenges? Some folks see AI as an attack, and some see it as a way to remove some drudgery while they focus on harder problems. It is about mindset.
>> It's hard to be positive about the idea of your skills getting devalued and getting kicked to the curb.
> I think it depends where people build their own identities in the value stream. Do you see yourself as.... It is about mindset.
No. You're missing the point in a pretty serious way. In large part, we're talking about lower levels on the hierarchy of needs than that.
> Some folks see AI as an attack, and some see it as a way to remove some drudgery while they focus on harder problems.
Some folks see it as a way to remove hard problems so they can focus on drudgery. Do you love to code, but hate code reviews? Guess what you get to do more of now!
> Building greenfield prototypes is now 100x easier!
And then the boss-man can use his MBA, fire 90% of the team, pocket their wages, and get 10x the prototypes. Repeat that through the economy. Your teammates have 20 years to retirement and can't pay their mortgage. "Progress!," but for whom?
I suppose you would need to weigh the utility points lost from people no longer enjoying their jobs with the positive utility points gained by the consumers of products and services.
My intuition is that with the law of leverage in mind, the former would be relatively low and the latter would be relatively high.
It is of course up to the government and culture to minimize the former and maximize the latter.
A ton of the reinforcement type training work really just aligning the vague commands a user would give to the same capability a model would produce with a much more flushed out prompt.
reply