I really don't understand how this can be. I have sat in fund raising meetings with VCs in toronto and my experience is that there is shit ton of due diligence at the tech level. a product which has no moat, was already available, was duplicated within a couple of days is valued at 7B - i thought we were past the peak of the hype cycle.
> sat in fund raising meetings with VCs in toronto
There's your problem. The single biggest thing every Canadian VC is trying to figure out is "why are these people asking us for money when if they were any good they'd be in the US" so by simply asking them you're already signalling something bad. A lot of their enthusiasm for process is based on this suspicion and also that the entire industry is just a way for various professional services to extract most of the investment money, since that's the game they're so used to playing with the government.
There are some Canadian VCs earnestly trying to improve but they are overwhelmingly hilariously conservative and focused on unimportant signals over reality. This is one (but not all) of the major factors that drive basically every remotely ambitious Canadian company to run a corp in Delaware and go for funding from the US. The tax situation is the other major contributor.
Canada doesn’t have throwing around money like in the US. We have resource extraction -> export money that’s it
There are boatloads of tax breaks and incentives that mitigate all the financial stuff and make running a startup here just fine honestly.
But that does nothing to make up for the terrible investment community. Getting started here requires already being started.
When I briefly worked for a Toronto startup, it was like all of them went to the same private boy's schools together as kids. It was a status club.
I jumped ship to an American startup and made almost double the money dealt with 0% of the bullshit and they were bought by Google the next year.
There is a belief that there is going to be at least one more breakout success in startup AI labs - rather than OpenAI and Anthropic being the final word - and so investors want to own a part of whichever companies seem most likely to be that success. If you start from that premise and stack rank what company that might be, you could quite reasonably put TypeSafe toward the top of that list right now, based on the people at the company and the ability they've demonstrated to ship stuff that people care about and cut through the noise in a crowded space.
Also the situation isn't static. Investors know that the act of writing them a $870M check itself increases the chance that they'll be one of the winners, because that will attract more talent, customers, and funding to the company in a self-reinforcing cycle. And investors know that other investors know that, and that someone is going to write them that $870M check, so to some extent they're forced to think of the company as having already been successful at the fundraising and already having that momentum boost.
Only a small number of investors in the world can play the game at this level, because you have to smart enough to be right (often enough), and you have to be established enough to see the deals (be on every CEO's short list - because CEOs are only going to seriously pitch 5-10 VCs on a hot deal, if that). Otherwise you can't pull it off. Martin Casado and his team are among the few that can and I think their results reflect that.
[dead]
The lack of a “moat” is mostly irrelevant because success is not decided by who can or can’t be cloned. TypeSafe invented[1] a new approach that became wildly popular almost immediately, if they can do that once, they can probably do it again. Venture capital is big bets, of course TypeSafe is going to fail, that’s inevitable, but if it has even a 10% chance of capturing 1/10th the market cap of OpenAI then it is a great investment! Plus, money means nothing any more, they’ve raised less at a lower valuation than Instinct, a personal assistant.
[1] not really but they did some innovative things and popularized a concept
I'd guess they justified the funding by revealing some grand scheme for a new product that they just need more runway to produce.
> was already available
no, it was not
> was duplicated within a couple of days
was it already available or did it become available in a couple of days? it cant be both (neither is true, actually)
The tech behind it existed. They made it a specific product, and got replicated in days.
Unclear what you're referring to. Please stop making vague claims and be specific.
maybe you entered the AI space during the vibecoding era but there had been ton of useful models before that. especially zero shot models- both for text and images.
Everything you said here is false. No, I didn't enter the AI space during the vibe coding era. I was training custom ML models back in 2017. And no, there haven't been models comparable to Jev before Jev was published.
Jev is:
- accurate
- general purpose
- fast and cheap
Models we had before Jev had at most 2/3 of above qualities, but none of them were 3/3.
What you're missing is: cost and speed. Otherwise, it's very much like running an LLM and constraining the output.
I don't know how good Jev's "confidence scores" are, but I would be surprised if they were in any sense better than logits from some good LLM. One advantage of Jev here is that the confidence scores are easy to access. Most LLM API providers don't provide an easy/convenient way to access the logits. But that's a minor point, you could of course build something like this with LLMs (and many people have).
llama.cpp is already extremely fast for single-token responses (<5 ms). I can't see Jev being faster when taking network latency into account, except maybe for multimodal inputs.
> Models we had before Jev had at most 2/3 of above qualities, but none of them were 3/3.
You're the one being deceptive here. Jev is trading accuracy, speed, and cost for generality. It's less accurate, slower and more expensive than trained classifiers. So it's still 2 out of 3, but with decimals. Maybe 2.2 out of 3 if I'm being charitable.
And the reason we didn't have that before is because nobody thought it's a good tradeoff.
When you say "trained classifiers", you are referring to models which are trained (or fine tuned) to work on one specific problem, right? That is the opposite of "general purpose".
Would Jev be more accurate in a specific task if it had been developed only for that task, as opposed to general purpose? Of course it would. So, sure, Jev is trading accuracy for generality. According to you "nobody thought it's a good tradeoff", which again is false, there was huge demand for a cheap and accurate general purpose classifier.
> work on one specific problem, right? That is the opposite of "general purpose".
A business doesn't need Jev for the sake of Jev. Most business are solving specific problems.
And fine-tuning got a lot cheaper these days - I've seen claims here on HN that ~500 examples is enough to beat Jev.
> "nobody thought it's a good tradeoff", which again is false, there was huge demand for a cheap and accurate general purpose classifier
There wasn't. The hope is that there was a latent demand, but we've yet to see if it's truly latent or just manufactured.
Noone is saying "hell yeah, finally we got a general purpose classifier, my business needed it so much". The typical message is "this seems cool, let me see where I can apply it".
The fact that name itself is a play on Jevons Paradox illustrates that there was no demand until Jev was released.
Yes, businesses are solving specific problems, but most businesses have more than 1 problem to solve. No, it is not economical to pay a data scientist to develop a custom model for each of your tiny problems. It is often much more economical to use a general purpose solution, like an LLM, or now, Jev.
At this point there's no need to pay a data scientist. You can literally ask Claude to do everything for you: extract real examples, classify them, post-train a model, and ship an API.
Now, it could be viable if your business has literally hundreds of problems thats require classification. I just haven't seen those.
I treat the fact that almost noone was doing that as evidence that decision models aren't that useful/groundbreaking. That, and the fact that every single demo I saw was either fake (e.g. playing games), contrived, or plain wrong (e.g. using Jev for compaction).
No, we're not at the point where you could ask Claude to do all of that, unless you have a super easy problem to begin with and/or you don't care about output quality. Feel free to link a counter example.
>>I was training custom ML models back in 2017
but you truly do sound like an angry 19 year old from your arguments.
- accurate - on what? on trust me bro benchmarks?
- zero-shot model are fundamentally general purpose.
- fast and cheap ; models on hf are FREE and fast enough.
"Models on hf" (unspecified) are "FREE"? Like "free to download"? Sure, but nobody was talking about that. They cost money to run inference on. Unless you are talking about some tiny toy models that are useless for any non-toy problems. You clearly don't have any idea what you're talking about. Just stop, man.
I think it is clear he is referring to zero shot classifiers with an LLM backbone. That tech has existed for a long time.
Its not normal time in SF/Bay Area. For better or worse, VCs in this city/region are thinking very differently on AI bets.
I think the key differentiator was that a team found a whitespace in what ChatGPT was doing, main comes from the same pedigree and team is as conscious of marketing as their product. SF VCs love these out of the box challengers, and people are claiming to replicate doesn't seem to matter.
The amount raised feels surprising but again entire SF/US AI scene is primarily "add moar layers and GPU" one trick ponies at this point.
This is the difference between a "hot" company in a good ecosystem like SF. Yeah, the funds can take 2-3 months to do their due diligence. By the time it's done, the round has closed, and then what good is the due diligence?
Because every VC knows that one of OpenAI/Anthropic/Nvidia/Microsoft/Google/Meta/SpaceXAI/AMD/Stripe... will acquire them within the next year for talent alone.
They got money and it needs to be put to work!
Probably the “nobody ever got fired for buying IBM” effect. If there’s a use case for the tech, buying the most well known implementation of it will always be useful for people who want credit without the threat of blame. This funding is based entirely on the hype and a bet that TypeSafe will have name recognition.
Isn't this the exact opposite? When people were saying that saying, the connotation was that IBM was an old, stodgy company that had been around forever. (These days I often think "No one got fired for choosing AWS"). Typesafe is a hot new startup that could, to my eyes, easily burst into flame or die in the next year.
I’m saying the bet is in them becoming THE System One Model company.
You are used to dealing with companies where the money bags hold the power.
When you are in the middle of a boom cycle, it's the hottest company that has the advantage. Investing in them is a matter of privilege and they get to pick and choose.
Also, Canadian VCs are bottom of the barrel as far as VCs go.
It might be the people that are being acquired too, at huge inflated ai researchers salaries.
Acquired in the vc sense… not literal exit.
A very major part of Jev is the cost and speed. Yes, classification is/will be a commodity business, just like LLMs are, and similarly there is no moat only production cost and pricing.
Yes, anyone can wrap a decisions API around an LLM, but so what? If you want to compete then you need to compete on price, and it's not clear if OpenAI and/or Anthropic are able or willing to do that without building a custom architecture, and even then is a race to the bottom on pricing really what they want to pursue?
I'm not sure if OpenAI have announced pricing for their Decisions API, but they have said it's based on Luna which costs $0.10/M input, not even remotely competitive with Jev's $0.04/M input, which I'd expect has some headroom built into it.
Assuming that the architecture behind Jev is not just an LLM, and gives them some inherent efficiency/cost and speed advantage, then the question is whether OpenAI and Anthropic really want to duplicate this and have a race to the bottom on pricing for what may be a large part of the business automation market they are addressing. Is that what they want as their IPO pitch - we're selling potatoes, and think can grow them cheaper than Typesafe ?
> there is shit ton of due diligence at the tech level
Maybe in some cases. But counterexample, courtesy of The Information:
"It took just 15 minutes for Blue Owl executives to agree to invest up to $10 billion in future projects alongside real estate firm Primary Digital Infrastructure during their first in-person meeting two years ago, said Primary chief investment officer Bill Stein."
https://www.theinformation.com/articles/blue-owl-eyes-new-de...
AI seems to make some people lose their damned minds.
The API was duplicated, the results were not.
Yeah your problem and my problem and others around here is the word you just said there... "Toronto." Canadian investors are risk averse as hell. And cheap. They can make more money helping sell bitumen or real estate, why bother with arcane tech?
And if you could put the words "Bay Area" or "Stanford" or "San Francisco" next to your name... different story.
The VCs are not buying the idea or the tech, they're investing in the people. And they invest in a formula that has already worked for them before to make big coin. Prop somebody up, let them hire like crazy, and then get them get acquired, and then cash out. They don't care if it fails if they can make it succeed 1/200 times.
Canadian investors want you to have already succeeded before they help you succeed a tiny bit more.
That is not how it works in the US.
What was the moat of Dropbox? Of Instagram? Of GitHub? Or of countless other very successful startups when they started?
Anyone know when this "have no moat" meme appeared? Even 5 years ago I don't remember seeing it on every post.
Whether a product has network effects makes a difference.
I would argue Dropbox did have a moat. It didn't merely store your data. It made it possible to make backup efficiently when bandwidth wasn't all that good.
Reading "no moat" so often is also tied to the fact those companies happen to be getting surreal valuations, at a quite early stage, showing no profit, building a tech that doesn't seem difficult to reproduce.
Instagram clearly did have one - Zuck had to acquire after internal efforts failed.
Dropbox had a super smooth UX that somehow nobody else replicated (seriously Google wtf). Instagram and Github won on network effects.
When you frame it as if we're in the Pets.com era of AI the continued gold rush makes sense.
brb gonna wrap claude with a new form of prompting and raise 500 mil