A couple weeks ago I rebuilt the https://hiresynth.ai landing page and docs using vitepress and was super impressed with the speed of implementation and how pretty it was by default.
This feels like the start of a lawsuit. If Tesla is state of the art and requires 5x less funding per charger than competition; what is the reasoning behind picking the competitors?
Here’s a better question. Is Tesla leveraging their market cap and cash reserves to capture a monopoly position on chargers? I’d want to make sure that market competitiveness is a thing. Otherwise you have a company that turns its behavior on all fronts into rent seeking.
I don’t doubt their efficiency and scale effects stack up to an enormous advantage. Absolutely their ability to finance impacts the unit costs of chargers very few if any other company could match. However:
* being extraordinarily well positioned to compete isn’t a crime or legally anticompetitive. It might be if they had a non competitive market monopoly in another market they’re leveraging in this market. But they don’t.
* lose leading isn’t illegal or sanctionable unless it fits fairly narrow constraints in antitrust law
* probably most importantly they have the only proven ability to deliver a critical national infrastructure to meet national energy, environmental, and security goals. They will be afforded remarkable leeway for some time.
* they’re actually weakening their grip on the EV market by opening the charger adapter spec and licensing access to competitors. If the supercharger network were plausibly their money maker eyes might squint. But it’s not. It’s an enablement to the entire industry, including their competitors. And it’s not reasonable to say the EV market lacks competitors.
That doesn’t mean this won’t change in the future. But I fully expect a Standard Oil world where EV charging is owned by Tesla. Then, legal market remedies would be squarely on the table.
> being extraordinarily well positioned to compete isn’t a crime or legally anticompetitive
Actually, using profits in one sector to dominate a second is pretty classical monopolist behavior. I’m fine with competition but if the bid is insincere because they’re not going to do the work for that number - the winner typically pays what the 2nd bidder bid typically in such contracts - or because it’s to capture a monopoly position and recoup the rest of the investment through rent seeking over the long term, these are negative behaviors we should discourage through various means.
> And it’s not reasonable to say the EV market lacks competitors.
That was true of the internet too. Yet somehow we ended up with a handful of major tech companies.
> But I fully expect a Standard Oil world where EV charging is owned by Tesla. Then, legal market remedies would be squarely on the table.
Wait. So you agree this strategy that Tesla is employing is likely to cause a problem in the future but your position is to remedy it in the future after it’s obviously an even bigger problem with jobs tied to it making it more politically risky, the legal and political climate being a joke on antitrust at the moment?
Also the rest of what you wrote is basically “there’s no law against it”. I’m saying there should be laws against predatory pricing and simple ones like your penalty is paying a fine which is based on what a fair price would have been and how much you’ve cost the economy through that behavior which government regulators and academic economists get to determine independently (& pick the highest price). You don’t need to do this for all companies - just ones who have enough cash flow that such behaviors are possible in the first place (a startup that’s not turning over a billion or so per quarter has a hard time pulling this kind of shit). Oh and treat fundraising like revenue a bit in that it counts towards you being regulated (eg if you’re market cap in a private market is >1B you can’t have any loss leading products).
* You have to have a monopoly in the other sector for it run afoul. Many companies enter new segments by loss leading taking profits from their established non-monopoly business to subsidize their growth in another segment. That’s not illegal in the least and is extremely wide spread. In fact the opposite in this situation is true. The EV market is extremely competitive but the supercharger market is dominated by Tesla. The fact they’re opening it up for use by competitors in EV means they’re not leveraging their supercharger hegemony to kill off EV competitors.
* “the internet” isn’t a single market segment. There are a lot of monopolies though, and I would point you to the ongoing antitrust lawsuits.
* you can’t proactively take someone for violating a law in the future to court. Yes they’ll likely be standard oil given how things are playing out. But they are not now. They would rightfully complain they’re being punished for future behavior that they may not do, and there may very well be a large competitor that springs up along the way. The market is too new and the government doesn’t generally interfere in market dynamics in nascent markets; and they definitely don’t punish you for something that might happen in the future. It’s not minority report - you actually have to commit a crime to be legally culpable for committing a crime.
* there are lots of laws about pricing. But in a competitive market companies are allowed to loss lead. But to your point there a lot of ways that can happen that’s structurally unfair. Using one monopoly to predate in another market is potentially illegal. Competing on price in multiple competitive markets is not. And it shouldn’t be, because that’s almost always how companies break into new markets and gain market share against established competitors.
I’d note regardless in this situation I will wager they do have an enormous cost advantage over everyone without loss leading. They have a scaled industrialized operation that generates revenues. No one else does.
> Many companies enter new segments by loss leading taking profits from their established non-monopoly business to subsidize their growth in another segment. That’s not illegal in the least and is extremely wide spread
Again, you’re arguing that something is legal when I’m saying the law should be changed. There’s no economically beneficial reason to allow anyone to loss lead by taking profits from one space and using it to capture market share in another, regardless whether you’re a monopoly to start with or not. That’s literally one of the well established mechanisms for a monopoly to start with. And once they get entrenched, they repeat that pattern to protect themselves from competitors, lowering the price until new entrants leave.
There’s cases where loss leaders I’m more OK with, but that’s usually around when you’re selling multiple “fungible” items and you have a loss leader on one to drive traffic & it’s not the manufacturer of the item giving you a deal to make it a loss leader. Think grocery chains, outlets, etc.
> you can’t proactively take someone for violating a law in the future to court
No where did I propose this. All I suggested is that there’s enforced regulations to make sure that Tesla is behaving here in a way that won’t result in needing an anti trust case in the future. Microsoft’s anti trust lawsuit went nowhere. Sure it’s not as entrentched, but if they hadn’t built their monopoly position in the first place & stuffed their coffers, they wouldn’t have been able to survive the massively bad decisions they kept making for a decade or more.
You seem to be taking on faith here that this is purely because Tesla has leveraged vertical integration to get a 5x savings cost. I’m skeptical though - if that were the case they wouldn’t need to be going after government bids for charger networks in the first place & could undercut on price. There’s just no way vertical integration in the car manufacturing space is yielding such impressive dividends in the charge manufacturing / operating space.
I think the key is I think the law is largely well thought out in this space if often poorly implemented. There is a balance between exerting control over markets and letting markets run rough shod over us. Not every practice that’s distasteful need be controlled nor every whim fulfilled. Loss leading is a way for monopolies to be broken too - imagine a world where Tesla is well established with super chargers everywhere and has become complacent. It’s fat and inefficient, doesn’t use the best tech anymore, and rests on its market dominance. Startups just get crushed under foot. You could do some sort of antitrust thing then but that doesn’t ensure competitiveness. See Netscape vs MSFT, the action was way too late to be effective. It wasn’t until another adjacent company with a huge war chest loss led their way into the browser market that internet explorers illegally monopoly was actually broken up.
So in the Tesla case, imagine now CATL sees an opportunity to leverage their battery dominance to enter the charger network business. If they don’t loss lead they face insurmountable penetration challenges even with superior technology that integrates better with their batteries. The capital and operational overhang is too great to immediately be profitable with a starting from 0% market share. The margins would have to start at infinity and work down as share grows to not loss lead. No one would pay a near infinite amount of money to charge their car. So A way to gain market share quickly would be to subsidize the bootstrap of their market share by pricing way under the cost covering margin required, and probably one that’s much lower than Tesla and is more in line with their at scale margin in the future if they’re successful.
Is that fair? Maybe not in some very strict sense of the word. Does the same technique work in less established markets? Absolutely. What do you think startups are doing with VC money? Are the financiers of startups adorable small business owners trying to make it in a tough world of hard nosed business so deserve special carve outs? Because they’re the ones subsidizing the loss leading startups do to establish themselves. The startups are just the workers, the capital is coming from ugly megawealthy folks with near monopolistic control over capital deciding who wins and loses, seeking to create more and more monopolies from the winnings of their prior monopolies.
Edit: on Tesla’s efficiencies I actually wouldn’t be too surprised if their relative cost covering margin is closer to 5x than not. However, expansion requires enormous capital investment and charging networks even at maximized margins can’t compare to return on investment to car sales. It’s an enabler for car sales scale, and a rising tide for EV lifts all boats. Every car maker wants to sell cars before charging services for that reason. So if they have $10 of capital for car making or $10 for charging infrastructure the cars get the money. So why does Tesla want the government to subsidize capital investment? Because they would rather deploy their earned and fundraising capital on their core business. The government wants the infrastructure and frankly other car makers want it too. Tesla takes operational risk, covers a non trivial amount of the capital commitment, and everyone gets the charger infrastructure we need in the present. The optimization of the competitive market would come later - once there’s a market to optimize. Frankly at a certain point Tesla would almost certainly spin off their charger network as it become a commodity and a capital distraction. Other car makers might compete then when their margins in cars are saturated, but I would put my money on gas station chains and oil companies deciding to get out of the age of fire and getting with fundamental forces and maxwells equations.
> Is Tesla leveraging their market cap and cash reserves to capture a monopoly position on chargers?
No. They are simply the only company that takes charging seriously.
I was helping to go through DCFC manufacturers' bids in 2021, and most of them had a "large sunlight-readable screen" listed as a fucking advantage. Because you can use it to show freaking ads or "branding".
Because that's certainly what I want to do: stare at even more ads as the charger does the slllllooooowwww negotiation process.
This was in 2022 and at least the common Superchargers deployed now don’t accept credit card payments and only recently were opened up to other manufacturers. A municipality could choose another manufacturer that supports more vehicle models.
This would be a very large concern I would have about trying to enter into any kind of an agreement with him specifically, but it is a criteria considered by pretty much everyone. Will this person or anyone in leadership be a problem later? It has an effect on decisions when making long term contracts.
This feels as if they've tacked the bing "chat with gpt4" to chatgpt while the previous version was able to browse multiple places and provide (usually) decent responses.
Have you found any custom instructions to improve the way it browses and responds?
That doesn't really look right to me, it looks like that's for responding regarding uploaded documents. I see nothing related to infinite context.
Also this is the azure repo from OP, nothing to do with the actual ChatGPT front-end that was asked about. I highly doubt the official ChatGPT front-end uses langchain, for example.
I don't see anything related to an infinite context in there. There's only a reference to a server-side `summary` variable which suggests that there is a summary of previous posts which will get sent along with the question for context, as is to be expected. Nothing suggests an infinite context.
Google released PaLM pricing per character at $0.0005/1k (for chat, instruct is 0.001).
With OpenAI tokens at approx 4 chars per token, this matches $0.002/1k tokens for 3.5.
Let's hope OpenAI retaliate, cut prices to maintain market share and start a happy race to the bottom/affordable genAI for users.
In my experience it's not as good as even GPT-3.5 at following instructions (for example formatting a text blob into structured json) which leads it have too much variability in output to be useful.
Have recently been working with it as well and come to the same conclusion. Would be nice to use it since my company is already a GCP shop, but the json parsing is super frustrating. I have noticed using Python's ast.literal_eval helps, since it catches single quotes.
Would recommend for any text heavy static site.