If it ever feels weird - just watch Silicon Valley show again.
"Revenue? No no no no. Why would you go after revenue? If you show people revenue, they’ll ask ‘how much?’. And it will never be enough. The company that was the 100x-er or the 1000x-er becomes the 2x dog. But if you have no revenue, you can say you’re pre-revenue and you’re a potential pure play.”
We took it now from no revenue to no actual product, or even a concept of a product.
Right, but, to be clear, that's not what happened in the case of the original post. They claim they were a senior dev who got replaced by several less effective junior devs.
> Decentralized solar plus batteries is already cheaper than electricity + transmission for me at my home in the US. The only thing stopping me is the permitting hassle or the contractor hassle.
Does decentralized solar plus batteries give you same amount of reliability? How many days without sunny weather can you survive without having to change your energy use habits?
Each 9 of reliability for infrastructure is EXTREMELY expensive. And grid has a lot of 9s.
But having electricity 13 days every two weeks is much better than not having it at all.
This isn't about China building out their grid with an over capacity factor of 200% so they can keep everything running even if rain, sun and wind all fail for months on end. This is a developing county getting to the point they can charge mobile phones consistently.
When I go to https://model.energy/ and ask it to solve for providing steady output in China from 100% renewable energy (wind/solar/battery/hydrogen) at minimum cost using 2030 cost assumptions and 2011 weather data, the solar curtailment is just 7.3% (and most of the energy is coming from solar, not wind). If I remove hydrogen and solve again, solar curtailment increases to 16.7%. "200% overcapacity" is completely bogus.
Try that again with 99% renewable and it becomes much more reasonable with regards to over overcapacity. 1% non-renewable would be a very good outcome.
One can enable "dispatchable 1" which is simple cycle gas turbines, and limit the total CO2 emission so that's at most 1% of the generation. Doing that, and with no hydrogen, solar curtailment is reduced by more than half, to 8.1%.
And what was the storage requirement? I just ran those parameters myself with China's 2.9 TW of constant electricity demand, and the storage requirement was over 70,000 GWh of battery storage.
By comparison, global battery production is around 1,000 GWh per year.
Battery production capacity grows by 10x every five years. In 2021 there was ~100 GWh of batteries produced a year. In 2031, it's going to be 20-30TWh per year. Current batteries have 10+ year warranties, and last 20-25 years. We're likely to see 30 years+ for the newer sodium ion batteries.
For something like 20 years, people have been looking at the exponential growth in the annual solar deployments and saying "well that's it, starting next year we're only going to deploy exactly as much as last year, plus 5%-30%". And every year these predictions are proven wrong. And every year they do the same dumb thing again:
It was around 14 hours of battery storage. Seems reasonable.
Realize that replacing all ICE road vehicles in the US with 70 kWh BEVs would require storage equal to ~40 hours of our average grid usage. The future is going to need large numbers of batteries, which is why China has been all in on this.
14 hours of battery (~40 TWh for China) with the hydrogen storage or without? Because the calculator was reporting ~78,000 GWh battery storage with China's weather selected, and 2030 technology assumptions. I changed the spatial capacity factor from 1 to 2 and the battery storage requirement dropped down to 68 TWh, but still well above 40 TWH.
Regardless, 14 hours of China's electricity demand is a whopping 40,600 GWh. By comparison, 2024's lithium ion battery production figure was 1.5 TWh [1]. Even assuming 100% of this went to EV's we're still talking about roughly 25 years worth of global battery production to fulfill only China's demand for storage in this model. As you point out, we still have loads of battery demand for EV adoption, so nowhere near 100% of production will be able to be diverted to grid storage.
The scale of storage required to make intermittent sources viable without being backed by a dispatchable energy source really is tremendous, and this often gets overlooked in pushes for a fully renewable grid.
Battery production capacity grows by 10x every five years. It was four years ago when I first heard that, and we are exactly on track still. In 2031 we will be at 20-30 TWh/year production capacity.
There are few things that grow this fast when it comes to manufactured things, atoms are far harder to arrange and scale than bits. But it's happening at a tremendous scale. Natural gas turbine production capacity is tapped out with long order queues, and so is battery production well into 2026, but only battery production capacity is expanding at breakneck speed.
Understand that only ~6 TWh of lithium batteries have been produced to date. As in, every single year of production combined adds up to less than 6 Twh. Moore's law largely stemmed from the fact that making a processer faster also meant making transistors smaller. Reducing the width of a transistor to a half, a quarter, etc. increased compute per cm^2 by double, quadruple, etc. Chemistry doesn't obtain that kind of exponential growth - we have hard limits on the number of joules we can store per gram of anode and cathode, so scaling up production means digging up more anode and cathode material out of the ground. The nature of resource extraction is that the easiest-to-exploit reserves are exhausted first, and continued production is contingent on accessing the progressively more and more inconvenient reserves.
Maybe in 2030 annual global production will be 30 TWh - we'll know in 4 years. But there's a lot of people who probably don't want to make trillion-dollar investments gambling on that possibility panning out.
Regardless of your confidence in battery production's continued growth, I think you'd agree that if someone is making a calculation about the required amount of overproduction required to maintain a stable grid, they should at least mention that their calculation is contingent on provisioning tens of terawatt hours worth of grid storage.
Getting to 10x production capacity doesn't require improving battery tech, it just requires building more factories. The equivalent here isn't Moore's Law, it's fab capacity. If battery tech stalled out today at the same pricing (it won't), we could still 10x the battery production capacity every 5 years just on this pace.
The learning rate for batteries has not been as steep as Moore's Law for ICs. But the value of being able to store mass quantities of electricity at low cost is so incredibly valuable that it's going to blow up into a huuuuge number of factories.
You look at the 6TWh of all time and see that as a limitation. I haven't seen that stat before but I trust you, because with the growth rate of battery production it has to be a tiny number, because it's exponential growth. In 2024, 1.2 TWh of batteries were produced, 20% of all storage capacity ever. That was a single year! What if, in 2024, we produced 20% of all CPU capacity every produced? That's obviously a hugely growing market.
> The nature of resource extraction is that the easiest-to-exploit reserves are exhausted first, and continued production is contingent on accessing the progressively more and more inconvenient reserves. Maybe in 2030 annual global production will be 30 TWh - we'll know in 4 years. But there's a lot of people who probably don't want to make trillion-dollar investments gambling on that possibility panning out.
If you spend a small amount of time diving into the industry, you'll see that there's a massive number of very smart people solving all these resource constraint problems, securing supply chains years in advance, and building like fucking mad. Sure, there's a lot of people that don't want to get involved, but they will be left in the dust.
We are witnessing a massive energy interchange. This is like when the PC came along, but much bigger in terms of quantity and speed. Sure, there are those who are still skeptical of energy storage, 5-10 years after it became blazingly obvious that batteries are cheap and getting cheaper and will take over the enery world. But they are the same people who saw the iPhone and said "it will never catch up to my BlackBerry."
Electricity storage in batteries is a swiss-army knife for the grid that never existed until recently. We couldn't do time arbitrage, always had to match supply and demand instantaneously, across grids spanning hundreds and thousands of miles. No more, that's all gone. We can do tiny microgrids, we can do single houses, we can do 10 TWh installs across grids, because batteries scale small, scale big, are cheap, getting cheaper, and are being produced on a growing scale that most people do not understand.
That was about the amount in both cases. Slightly more in the no-hydrogen case than otherwise. Hydrogen contributed only marginally.
Yes, it's a lot of batteries. So what? It's not like the current battery production is some firm limit. If anything, the very large future demand ensures batteries will be driven down their experience curve, so the cost will be even lower than assumed.
The world spends something like $10T per year on energy. Any replacement energy system is going to be a big thing.
You need to make an argument that is more than you expressing fear of large numbers.
Unfortunately, all such calculations are egocentric. People assume that everyone can use solar panels for 13 days 2 weeks, and when needed, we’ll just get electricity from the grid. But what they don’t take into account is that when there’s load today but none tomorrow, the grid becomes unstable.
2) This also increases costs. You might save electricity consumption in 14 times, but your expenses for grid electricity can increase in 14 times, because the grid still needs to be maintained — staff must be kept at power plants to ensure you can be supplied with 100% of your energy at any moment.
The tricky thing in cold climates is the part of the year when solar power is lowest but electricity use, for heating, is highest. Sometimes they have hydro or something.
Sounds good until you try to run a business. Having businesses randomly out of commission is not a way to bring country from developing to developed status.
Even if you have an under-provisioned solar+storage solution and don't want to splurge for a generator, even on cloudy days you still get power, just less.
Generally businesses are really great at balancing costs, and for highly-cost-constrained businesses if you give them 95% uptime at half the cost, the equation becomes clear. And in Africa, if the option is 95% uptime or 0% uptime, the choice is even clearer.
Where I live, I only get two 9s from the utility. And I'm within commuting distance of Seattle. With my generator, I still got three nines the one year where the battery tender failed and the generator didn't start when needed, but only because that outage was less than 8 hours and I replaced the battery tender before further outages (I could have jump started the generator, but the outage started overnight and waiting it out was easier). Most years, the number of brief outages adds up, and I probably only get five 9s.
Solar + battery + generator for really bad weeks (but make sure you exercise it!) could pretty easily add up to the two nines I'd get from the utility here.
For developing countries, solar + battery alone is likely be better than many grids which often are intermittent rather than 24/7 and many places don't have any access to utility power.
I actually counted the number of outages after I got my battery unit in June -- it was six in five weeks, for anything from a couple of seconds to 30 minutes, which I noticed because the unit clicked over to running from the battery, and the clock on the oven (which is still only mains powered) flashes until I go over and hit a button.
In April I had a 40 hour outage after a storm. That's what caused me to order the brand new Pecron E3600LFP, first New Zealand model shipment in "early" June (I received mine June 19).
In February 2023 I had a 4 day outage during/after a storm.
There are even, every 2 or 3 months, scheduled and notified 9 AM - 3 PM outages for equipment maintenance, tree trimming etc. Just those alone lower the grid reliability to around 99.5%.
Six days outage in three years -- let's call it four -- drops grid reliability by another 0.4%.
So, yeah, two 9s is about right.
With the Pecron base unit (US$999 at the moment still on Halloween special, $1259 before that) I simply don't notice any outage under 4 hours, and that's even with a full winter heating load. In fact I deliberately turn the mains to it off from 7-9 AM and 5-9 PM every day.
A 4 hour outage was a little close sometimes, so in August I added a 3kWh expansion battery ($699 on pecron.com right now).
With 6kWh I can run my fridge, computers, Starlink, some LED lighting for 36 hours. Or 30 hours with typical kitchen appliance usage added (espresso machine, toaster, kettle, microwave, air fryer).
Or virtually forever now I added 6x 440W solar panels (cost me US$400 total) to it, which still generates around 200W between them in even the worse overcast and rain.
I'm running this stuff as a mini off-grid system, not connected to the house wiring at all -- except plugged into a standard socket to charge the battery if needed. I also have a $450 2kW petrol generator which I can use to charge the battery if needed, but needing that should be very rare.
Total cost: under US$3k. More like $2.3k at the current Halloween special prices.
The grid has a lot of 9s, but in a lot of places losing power for a day or two after a storm is not unusual at all. The grid per se being fine but your actual neighborhood being dark for a couple days is a pretty common experience in some places.
Last time my building lost power was about 19 years ago, when I was living in a Welsh valley halfway between the two nearest villages.
Since then, none of the extended Portsmouth conurbation, Sheffield, Cambridge, rural Cambridgeshire or Berlin have had any problems big enough to even notice while I've lived in them.
I have seen at least two circuit breakers trip in that time though.
In the UK, I think I can remember 3 power outages my entire life. One when there was significant flooding in my hometown as a child, which lasted around a day, once at university for a few hours (local substation failed) and recently 30 minutes overnight while they were upgrading something (with a lot of notice). I may be undercounting/misremembering but I don't think its far off.
I think the main difference is the UK in all but mostly rural areas has all the power lines underground. This is very different in eg North America where you can go a few blocks out of downtown areas and it is all overhead delivery.
I don’t know where you live, but I experienced outage in Budapest once in at least 10 years while I lived there. And only one phase was out, not all. We even lamented with my friends that we didn’t even remember when was the last time when something like that happened. I never had to reconfigure the clock on my microwave, just for daylight saving time. I know that even 30 kms from there my granddad still experiences outages monthly, but there are places where that happens very-very rarely nowadays.
And, for a refrigerator and a lot of loads, being down for 2 days straight is way worse than a few hours a year. losing 48 hours of supply a year if broken into 2 hour chunks is not nearly as bad as losing 48 consecutive hours.
I get your point, but I personally would be grumpy if I lost power for two hours twice a month. I realize that is rich considering this article is about people who are lucky to get any amount of power reliably
When I lived in a city proper, the grid was doing well to maintain 98% uptime. Multiple day long outages were the rule, not uncommon to lose power 3-5 days in a row.
Now I live in a rural area and it's uncommon to avoid outages more than a month. We have an automatic transfer switch and fuel generator from previous owners and it saves hundreds of dollars in frozen food.
This is in the US by the way. If you're investing in a transfer switch and generator now, the cost is going to quickly approach a modest solar + battery set up with a whole house inverter, and of course, you save money all year that way, not just in outages.
The grid where I live has a target of 89 minutes of unplanned power outages per year for urban customers and somewhere in the high 200s for long rural feeder lines. This is in Australia, where serving outlying customers comes with some geographical challenges. I think it's currently sitting at 99.998% reliability. I can't remember the last unplanned outage longer than a couple of minutes, although they did some planned work last year and took out our power for half an hour.
I'm surprised that someone would think days of power outages are normal everywhere. My family used to get hit with 8+ hour outages every few years back in the 90s because we were at the end of a single long rural feeder line, and we thought that was an unacceptable frequency.
I lost power for 10h in my city recently and it was a big fucking deal. The last 5 years that's the first time that happened. I would say I have less than a hour of downtime per year in the other years
Can confirm. I live in a US city and the only 9 involved is maybe the very first number. I've lived here just over a year and we've had 1 full day without power and probably 8 to 10 short outages between a few seconds and several 10s of minutes. I'm adding batteries and solar permitting be damned.
Wild! I’ve lived in Chicago and San Francisco and have never lost power for more than an hour. And can’t remember the last time it went out at all, maybe 2 years ago?
I'm (not GP) in the Chicago burbs and expect to lose power 1-3 times a year. Usually it's for less than twelve hours but last year it was out for three days straight. Most recent outage was ~10 minutes long a couple weeks ago - I still haven't set the oven clock.
The cause around here is usually storm + trees + above ground power lines, plus a low enough population density that you're not top priority for the utility company.
My concern is that it deflates any impetus to actually solve the problems of regulatory capture, profiteering, and other corruption.
Not everybody can afford the up front costs of installing solar + battery storage, plus replacement when the PV cells and batteries inevitably reach EoL. These people will be left behind on a decaying grid nobody with political capital wants to fix or at the mercy of landlords.
I really don't like this attitude we have in America where we realize "thing is broken" and advocate throwing it away instead of trying to fix it.
> I really don't like this attitude we have in America where we realize "thing is broken" and advocate throwing it away instead of trying to fix it.
Because people are too busy playing Team Politics instead of solving issues that everyone can get behind.
Fixing the power grid is one of those things that everyone could get behind, and yeah I agree, it disproportionately affects the economically disadvantaged.
It is funny to me how fractally perverse systems gets when a centralised authority refuses to directly solve a problem but rather decide have it solved by third party uncooperative players by creating an endless stream of byzantine rules to force the solution to be a twisted copy of what the centralised authority could have done by itself.
Of course there are failure modes in any approach but "oh no! Herding cats is hard. Who could have imagined!" is funny to me
According to PVWatts, a 10kW solar system would get me very close to my average usage in December. I'd be way over in the summer, could probably get away with a 4kW system and dial back use during an outage. I can lease two Powerwall 3 batteries from my utility company for $55/mo.
We used to lose power 3-4 days a winter in our old house. It would have been really nice to have heat. A generator or smaller system could handle that.
keep in mind the limitations of these forecasting calculations. On an AVERAGE day, assuming AVERAGE weather, assuming AVERAGE load, you should be fine.
The trouble with relying on the weather for your electricity is that it is entirely possible that you will go five days straight with cloud cover, limited to no solar generation and then be freezing. This is the problem that the electricity grid solves with varied sources of generation.
I read a decent essay about the difference between solar and wind reliability and fossil fuel reliability.
Solar and wind tend to be regularly and predictably intermittent but not unreliable. That's something you can design around. Especially when you have cheap storage to handle critical loads.
It's instructive to look at California's ISO website's supply graphs over the year. Renewables follow a reliable daily cycle.
Distributed can do redundancy. It’s relatively cheap.
Consider a family with two cars instead of one. How often do they have zero working cars? The correlated failure rate squares while the cost doubles.
My home now has a grid connection, house battery and solar, a caravan with mounted solar/battery/fridge/inverter beside it, and I also have a portable “powerstation” and portable solar panel which is basically a UPS. My fridge contents and phone charging needs have a several extra 9’s now for costs that have scaled very well.
These systems are tech that is improving rapidly. In some years these African farmers with their increased yields will likely add a bigger, second solar & battery system. In a village you can run a cable next door. Etc.
I mean, it very much depends on where you are. Three 9s would be no more than about 8 hours downtime per year. A lot of rural locations would do worse than that, realistically.
False premise. The company isn't demand limited, despite the (conspicuously implied and never actually stated) conclusion the headline desperately wanted you to reach.
I really wish people had any media literacy left. This brand of lying without lying is extremely common in modern media, and also extremely easy to spot once you know what to look for.
> KBA said Tesla sold 750 cars in Germany in October, down by 53.5% from a year earlier. The number of Teslas sold in the January-October period dropped 50.4% to 15,595 units, compared with the same period last year.
To be clear, you are suggesting that Tesla had no delivery "wave" between January and October? And that is the sign of a healthy company?
You fell for the "we didn't say it so technically we didn't lie" clickbait headline, I see.
This article is the same recycled misinformation that's been repeated for years. What's actually happening is that Tesla does regional delivery waves, which results in large month-to-month fluctuations. Nothing new here.
Yes Virginia, the media will distort information to sell eyeballs. Color me shocked!
I'm a hobby woodworker - I've tried using gemini recently for an advice on how to make some tricky cuts.
If I'd follow any of the suggestions I'd probably be in ER. Even after me pointing out issues and asking it to improve - it'd come up with more and more sophistical ways of doing same fundamentally dangerous actions.
LLMs are AMAZING tools, but they are just that - tools. There's no actual intelligence there. And the confidence with which they spew dangerous BS is stunning.
You do bare minimum, get sued, litigate for few years, pay some fines (that are MUCH less than amount of money they made), then do bare minimum again, etc - rinse and repeat.
And one may argue - it's not malicious compliance. It's just how world is setup for big companies. Any other strategy is leaving huge amount of money on the table - that's not what shareholders want. Unless you're willing to have fines that exceed the profits (and not just say about possibility of those, but actually impose and collect them) it's not going to change.
The fair use test (in US copyright law) is a 4 part test under which impact on the market for the original work is one of 4 parts. Notably, just because a use has massively detrimental harms to a work's market does not in and of itself constitute a copyright violation. And it couldn't be any other way. Imagine if you could be sued for copyright infringement for using a work to criticize that work or the author of that work if the author could prove that your criticism hurt their sales. Imagine if you could be sued for copyright infringement because you wrote a better song or book on the same themes as a previous creator after seeing their work and deciding you could do it better.
Perhaps famously, emulators very clearly and objectively impact the market for a game consoles and computers and yet they are also considered fair use under US copyright law.
No one part of the 4 part test is more important than the others. And so far in the US, training and using an LLM has been ruled by the courts to be fair use so long as the materials used in the training were obtained legally.
> And so far in the US, training and using an LLM has been ruled by the courts to be fair use so long as the materials used in the training were obtained legally.
Just like OpenAI is rightfully upset if their LLM output is used to train a competitor’s model and might seek to restrict it contractually, publishers too may soon have EULAs just for reading their books.
OpenAI's hypocrisy on this matter is precisely why hackers should be taking this as the best opportunity we've had in decades to scale back the massive expansions that Disney et al have managed to place on copyright. But instead of taking advantage of the fact that for once someone with funding and money can go toe to toe with the big publishers and that in doing so they will be hoist on their own petard, a lot of hackers appear to be circling the wagons and suddenly finding that they think this whole "IP" thing is good actually and maybe we should make copyright even stronger.
Surely making copyright even stronger (and even expanding it to cover style as some have argued in response to the Ghibli style stuff) will have no unintended consequences going forward into a future where more and more technology is locked down by major manufacturers with a strong incentive to use and abuse IP law to prevent competition and open alternatives... right?
That's certainly an argument often made about counterfeit goods, and it can certainly be true in cases (and counterfeiting has other problems, namely confusing the origin of a specific good when that matters to the consumer), but it's also not a universal truth either. Were it a universal truth, that would imply generally that open source can't work because anyone can make and distribute copies of the open material, but also it implies that Windows and macOS should not exist because of all the innumerable Linux clones.
Also instructive would be the IBM BIOS clone, it is perhaps true that the "IBM Compatibles" killed the market that existed for IBM machines at that moment in time, but it's also true that it opened whole new markets, both to the clone makers and the ancillary businesses, but also arguably IBM themselves.
3d printing and Arduino are probably other examples where "counterfeits" might have shrunk the market for the originals (Prusa is notably reducing how open their designs are, and Arduino themselves are not the healthiest, modulo being owned by Qualcom now), but the market for Aruduino projects and ancillary supplies and certainly the market for 3d printers is massively healthy, and arguably both are healthier than if Arduino or Prusa (or really Reprap) were the single and sole providers of their products.
And I think art has an even stronger bulwark in that a lot of the value of a given "art" comes not from the art itself, but from the artist. It's very possible many famous artist's works were actually made by their apprentices, but until someone proves that, the art will continue to have value as an original work of the artist. But art is also a dime a dozen (or less). The internet is full of free or dirt cheap art and today you can go on fiver or mechanical turk and commission any number of artworks for probably less than your day's wages. But no one is buying tickets to your Fiver concert. No one buys $1k per plate dinners at Deviant Art gallery showings. But they will pay many thousands of dollars for a piece of artwork that might destroy itself because the person who produced that artwork is named Banksy.
I don't think they are rightfully upset at all. Yeah, no kidding. Everyone becomes pro rent seeker when it suits them. Which is the exact reason we must rain it in
* The use is commercial (a number of emulators are paid access, and the emulator case that carved out the biggest fair use space for them was Connectix Virtual Game Station a very explicitly commercial product)
* The nature of the work is imaginative and creative. No one can argue games and game consoles aren't imaginative and creative works.
* Quantity of use. A perfect emulator must replicate 100% of the functionality of the system being emulated, often times including bios functionality.
* Impact on market. Emulators are very clearly in direct competition with the products they emulate. This was one of Sony's big arguments against VGS. But also just look around at the officially licensed mini-retro consoles like the ones put out by Nintento, Sony and Atari. Those retro consoles are very clearly competing with emulators in the retro space and their sales were unquestionably affected by the existence of those emulators. Royalty avoidance is also in play here since no emulator that I know of pays licensing fees to Nintendo or Sony.
So are emulators a violation of copyright? If not, what is the substantial difference here? An emulator can duplicate a copyrighted work exactly, and in fact is explicitly intended to do so (yes, you can claim its about the homebrew scene, and you can look at any tutorial on setting up these systems on youtube to see that's clearly not what people want to do with them). Most of the AI systems are specifically programmed to not output copyrighted works exactly. Imagine a world where emulators had hash codes for all the known retail roms and refused to play them. That's what AI systems try to do.
Just because you have enumerated the 4 points and given 1 word pithy arguments for something illegal happening does not mean that it is. Judge Alsup laid out a pretty clear line of reasoning for why he reached the decision he did, with a number of supporting examples [1]. It's only 32 pages, and a relatively easy read. He's also the same judge that presided over the Oracle v. Google cases that found Google's use of the java APIs to be fair use despite that also meeting all 4 of your descriptions. Given that, you'll forgive me if I find his reasoning a bit more persuasive than your 52 word assertion that something illegal is happening.
Well they are completely different systems functioning in completely different ways and only looking at one of the four factors isn't doing any favors.
I believe we’re in violent agreement here, because my point was that all 4 aspects are equally important and they need to be evaluated as a whole. And further that the current legal rulings on these systems delve into each of those parts with much more nuance and care than the provided 56 word surface level examination of the issues
It seems like you're responding to a question about training by talking about inference. If you train an LLM because you want to use it to do sentiment analysis to flag social media posts for human review, or Facebook trains one and publishes it and others use it for something like that, how is that doing anything to the market for the original work? For that matter, if you trained an LLM and then ran out of money without ever using it for anything, how would that? It should be pretty obvious that the training isn't the part that's doing anything there.
And then for inference, wouldn't it depend on what you're actually using it for? If you're doing sentiment analysis, that's very different than if you're creating an unlicensed Harry Potter sequel that you expect to run in theaters and sell tickets. But conversely, just because it can produce a character from Harry Potter doesn't mean that couldn't be fair use either. What if it's being used for criticism or parody or any of the other typical instances of fair use?
The trouble is there's no automated way to make a fair use determination, and it really depends on what the user is doing with it, but the media companies are looking for some hook to go after the AI companies who are providing a general purpose tool instead of the subset of their "can't get blood from a stone" customers who are using that tool for some infringing purpose.
reply