Reading through the memo, the GTM/use cases/market seem to be void of any kind of beachhead strategy?
I see lots of technical stuff, but the complete business plan & roadmap seems to be lacking any kind of tangible use case that is showing me how they plan to differentiate? (The technical reasons they state seem to be fairy minimal, and I don't see anything that competitors couldn't do either.)
The reason that I'm weary about this: I've spent a few years in an "innovation division" after customers repeatedly asked for such a service. Turns out the budget for long-running innovation-as-a-service is a hard sell, because in the end someone always has to pay for it, and innovation just for the sake of it gets old really quick. The long term survivors where the ones who continuously innovated for a particular use case - so more on a project basis...
So I'd assume that, just like with blockchain & VR/AR, the biggest value will be in very specific use cases in very specific verticals, as opposed to the "spray & pray" approach, offering an "AI as a commodity".
Or is this really just a EU FOMO play, i.e. US is doing it so Europe can't stay behind?
(Probably hoping for EU funds/grants via that "Cedric O" guy.)
I've read it like this: They differentiate themselves from OpenAI/Anthropic/other major players by being in the EU and (more) open-source; they differentiate themselves from other European competitors by being as good as OpenAI.
I think it's a fine proposition given the early market and their team, as long as this hypothesis holds true for years to come,
> we believe that most of the value in the emerging generative AI market will be located in the hard-to-make technology, i.e. the generative models themselves,
If you read the memo carefully, they do claim to be open-source at the beginning, but then switch to explaining that they will make money off closed-source models for their customers.
So, not more open-source than what OpenAI is doing.
Ok, so they're aiming for a more "spray & pray" kind of GTM, challenging the incumbents based on the premise that it'll be EU based.
This basically then boils down to: "ChatGPT got x users in a week, and we want to do the same for EU.", without a clear path to profitability or guarantees.
So it's effectively a humongously risky bet by EU VCs?
I think that the investors are mainly betting on the team expertise to train models as good as the one of OpenAI but with an open source approach (the main differentiator), which according to the founders would open up a bigger market opportunity for them.
> So it's effectively a humongously risky bet by EU VCs?
isn't it supposed to be VCs job to bet on risk proposition with the hope of a 100x return?
VCs keep lecturing candidates on making sure that they have paying customers on the radar from day one, so no, that does not look like a reasonable VC bet.
For me it is clear that they have given that much money expecting they can get back way more having access to EU bureaucracy.
In the EU bureaucracy is huge and they spend billions on taxpayer money. It really doesn't matter if those billions are wasted away or not.
I have a company and I have to talk with EU politicians, most of them are clueless regarding technology, but they spend billions anyway, so you need to talk with them like they are 10 years old.
It seems like they have a guy who is an specialist on just that and knows the right people.
Since they didn't even bother to create a deck, it seems to me that they were promised this investment already, and making the pitch was just a formality. "Create a company and we'll give you 100m funding", is how I imagine it went.
Second, seems like all six founding members are French, so this is really a French company, not a European one. Other countries will need to create their own becwof this.
> In the EU bureaucracy is huge and they spend billions on taxpayer money.
This is a fairly popular belief, but the EU's bureaucracy is actually pretty small. The entire EU civil service is only 40,000 people, most of them part time. Ireland is an EU country with about 1% the population of the EU, and also has about 40,000 civil servants, and 300,000 people in the wider public service.
1% of the per-capita civil servant requirements of an actual country really seems very lightweight.
(The US federal civil service is 2.8m people, but this isn't a great comparison, as it handles a lot of stuff that in the EU would be the domain on national civil/public service)
What we mean by "EU democracy" is all the medium to large-size companies that will suffer tremendous pressure from their respective governments to use an EU provider. Banks, industrial OEMs, etc. Lots of customers in perspective.
"Oh but non-tech people don't know the value of technology" Being condescending doesn't help much, and I bet they know a lot of things the tech people don't know or don't care about.
And you might even disagree why it's not important but it is good to be aware, and not simply dismissive
They wield some authority (their respective positions at big AI companies/research centers) in the hot topic of the hour and got a lot of money from the usual VC crowd. I don't think this "deck" shows how talented or whatever Mistral is, it just shows how stupid the VC space is. Remember, nobody checked the FTX books.
The Fed needs to raise rates significantly higher until this stops.
Actually seems pretty strong and I really like that it's a memo and not a deck ( I guess that's the Amazonian in me speaking). Fundamentally they are aiming for billion dollar valuations and the company and the initial investors want to start strong without the need for subsequent raises anytime soon .... so it might just work out!
They anticipated the need for a Series A in Q3 24 in the memo "In
that round (Q3 2024), we expect to need to raise 200M, in order to train models exceeding GPT-4 capacities."
It's a very impressive memo, but I'm not sure if I see anything here that's replicatable. It's clearly a sick team, whom have clients lined up (many of whom probably wont ever engage with an American AI firm until they find ways around Schrems II and, possibly III). Most four-week start ups are not in such a situation.
Mistral is getting a fair amount of bad press and unfair criticism. As an EU citizen, I am glad to see this initiative.
I am worried by a couple things, however:
- their market is not EU, but France: soon enough, UK and Germany will have their own Mistral. Each will have trivial access to their country's market (if only through politic backing) but accessing the neighbor's market will be an uphill battle.
- they are just helping EU play catch-up, not be at the forefront. Like sama said in a recent tweet: "the most special thing about openai is the culture of repeatable innovation. it is relatively easy to copy something; it’s hard to do something for the first time. it’s really hard to do many things for the first time!" [1]
> Seems to me that it's the team + market that won the money.
That's pretty much it. Most analysts expect there's an addressable market of $20bn this year, growing to $100bn per year by the end of the decade. The market will be split between four or five companies. The number of people globally with the required expertise to create these models is in the tens (generously one hundred individuals). Three of those people are involved in this company now, and another five are joining in the next few weeks, each leaving senior positions at competitors.
This one company has secured something in the order of 5-10% of the global talent pool for a $100bn/yr industry.
I don't at all believe the people able to build models are that few. Aren't tons of people doing open source models already? I believe the number of people who have already built models to be low, but the number of people with the capability to build models to be very large. It's not magic.
1. The founding team’s track record and credentials.
2. there is currently no major independent player in the EU of OpenAI’s scale.
3. EU entities will prefer working with an EU based partner over OpenAI.
Having the LLaMA tech lead is a win. Having someone ex-OpenAI can bolster their position even more.
Now the anti-thesis:
1. It is gradually becoming apparent that models trained on high quality data don’t need to be so big.
2. phi-1 from Microsoft is < 1.3B parameters and can compete with GPT-3.5 reasonably well.
3. In addition, it seems GPT-4 is not one model, but a collection of 8 models trained on different datasets. Along with the above, all of a sudden - training GPT-4 like models is coming within the reach of an average person.
All in - nobody has won this yet. And Meta open-sourcing their research was the cleverest strategic play that I’ve seen in a while. It has done so much to weaken their competitors while really pushing AI forward in an open way (so much for “OpenAI”).
It's not the pitch that raised the money, it was the hype around GPT-4, and the realization/fear that the EU could miss the next tech boom if they are inattentive.
To me Mistral is a cheap kitchen appliance brand so I looked it up. There’s a lot of brands named Mistral. Soap, windsurfing, apparel, alcohol, footwear, to name the first half of the first page of Google results. We can add AI to the pile. :)
I’m not sure I understand the motivation behind picking a well-used name. Is there some strategic advantage to this?
While I fully support a french effort to be competitive in the AI industry, the memo is mostly empty apart for the 1 line bio of the team members that could or could not show the seriousness of the project. I hope that it's just the "executive" summary and that there is a real, detailled memo behind the $100 millions and that it's not just another "let's fund our friends" from the local capitalist class
> We believe that most of the value in the emerging generative AI market will be located in the hard-to-make technology, i.e. the generative models themselves.
This may be true, but it will be a lot easier to target usability and niche use cases than the modeling...
> Cédric O: Former French Secretary of State for Digital Affairs
That explains a lot actually, France is the epitome of crony capitalism, and having a former minister (from the current ruling party) is going to help a lot here…
Cédric O is part of the wave of businessmen who went into politics recently to follow Macron. He's however a typical French industry guy : went to one of a handful of elite highschools, then went to a business school that is part of the French equivalent of the Ivy league, then went to work for former state-owned entreprises and conglomerates, and major suppliers of the French government...
In short, even before he went into politics, he was well-introduced in every clique of the French political-industrial milieu.
And (un)surprisingly he's one of the more modern industry guys.
> Cédric O is part of the wave of businessmen who went into politics recently to follow Macron.
Recently? He's been doing politics since at least 2006 (campaigning for Dominic Strauss Khan, a famous French politician better known abroad for his rape of a maid in a hotel in NY in 2011).
> He's however a typical French industry guy : went to one of a handful of elite highschools, then went to a business school that is part of the French equivalent of the Ivy league, then went to work for former state-owned entreprises and conglomerates, and major suppliers of the French government
That's true, but you're missing the key part here, which is what makes France special in that regard: he did that in parallel (and with the help) of his political career. He joined Safran thanks to his former MP Pierre Moscovici (Cedric O was his legislative assistant) who had become minister of the Economy in 2012 after Hollande's election (And Cedric O was the deputy director of François Hollande's victorious presidential campaign).
This is very typical of the French Elite, were people involved in political parties often go back and forth between the public and private (or semi-private) sector with the help of their political network. The current prime minister has the same kind of background for instance. And there's even a word for that in French, it's called pantouflage.
Yeah, because the Military-Industrial Complex in America is a purely competitive market, driven exclusively by the invisible hand. :-)
I find it very funny when Americans say things like that (not that France is not like described) without even pausing to realize we are a country that legalized and institutionalized lobbying.
It's a surprise to no one that the US politics is a shit show, but for this particular issue I think it's much more pronounced in French politics than in the US: We have both our President and Prime minister who've been doing back-and-forth between politics and executive positions in the private sector, and a handful of ministers (Catherine Colonna, Stanislas Guerini, Agnès Pannier-Runacher, Amélie Oudéa-Castéra) are in the same boat.
I didn't notice that line, but my immediate thought on reading the memo was that they were going hard after European capital. Seemingly both governmental and from VC firms in Europe that are well-connected to the traditional establishment.
(Even if they don't get government funding, it looks like they've already positioned themselves to get a ton of support from the EU and from European national governments, who want local champions to compete with the American capitalists.)
My thoughts:
- it's a fairly well-written memo: clear, concise, and relatively well-structured. It's a refreshing alternative to the endless series of pitch decks which mainly induce groupthink.
- the insights aren't that original. I would guess that they are strongest on the technical level. Their product and industry insights are less original or penetrating (pushing open-source and the European angle may give them some advantages -- it almost certainly will not give them an edge over OpenAI, who seem to have a relentless focus on iterating their product, a "live player" CEO in sama, and a strong dose of healthy scrappiness and creativity.)
- my prediction: they will stick around for a time, build something impressive with their technical team, raise more funding, hire a large and strong technical team (with their enormous warchest, status as the European champion, and appeal to the open-source community), and basically become a solid "fast follower". However, I think it's less certain whether they'll be able to do anything really creative or innovative.
Their overwhelmingly strongest market is going to be large European enterprises (or maybe governments), who will have lots of specific niche problems. The downside of that is their org then moulds itself to fit the needs of enterprise customers.
More speculative: the only way they can really become dangerous will be by attracting a bunch of young top-notch 18-25-year-old devs, ML people, or other builders (basically, people with creative energy and without commitment to the "established ways of doing things"), throwing them in a room together, and giving them a remit to come up with the most crazy, ambitious, unexpected applications of AI that they can. A universal translation app customised for the needs of ERASMUS students. Some weird generative music app for Berlin nightclubs. The actual ideas don't matter, what matters is having 50-100 hyperactive GenZ builders energetically exploring the space of combinations of technology, product and market to identify the genuinely novel applications of AI.
But I predict that they will not do that. You cannot cultivate that kind of free-spirited entrepreneurial DNA while simultaneously seeking deep connections with the European political establishment.
So, my overall prediction: they will focus on the underlying technology and on connections with governments/large enterprises, and deemphasise product innovation. Basically, building an AI-powered French IBM, rather than Google+++++. (Google+++++ will be built by one or two lone young researchers following their noses and working on weird esoteric ideas.) If I am right, and product innovation is the key to harnessing the power of generative models, then these guys in Paris will be formidable but not dangerous.
"We will distribute tools to leverage the power of these
white box models, and create a developer community around our trademark. This
approach is an ideological differentiator from OpenAI; it is a very strong point
argument for hiring highly coveted top researchers, and a strong accelerator
for development, as it will open the doors for many downstream applications from
motivated hackers."
So they're trying to build a community of inventive 3rd party developers, however:
"Business development will start in parallel to the first model family development, using the
following strategy
- Focus on exploring needs with a large industrial actors, accompanied by third-party
integrators that will be given full access to our best (non-open source) model"
Suggests they be will sucked into the enterprise market as I said above. Still, the last line:
"Co-design of products with a few small emerging partners that focuses on
generative-AI products."
Suggests a wildcard: one of these "small emerging partners" might come up with the $trillion black-swan application of AI.
[I'm going to bookmark this comment and reread it 2 years from now to see how wrong I was]
Who is going to be first to implement real time video analysis of politicians giving speeches and determining the truthfulness of what they say and growing their nose longer the more lies they tell.
I see lots of technical stuff, but the complete business plan & roadmap seems to be lacking any kind of tangible use case that is showing me how they plan to differentiate? (The technical reasons they state seem to be fairy minimal, and I don't see anything that competitors couldn't do either.)
The reason that I'm weary about this: I've spent a few years in an "innovation division" after customers repeatedly asked for such a service. Turns out the budget for long-running innovation-as-a-service is a hard sell, because in the end someone always has to pay for it, and innovation just for the sake of it gets old really quick. The long term survivors where the ones who continuously innovated for a particular use case - so more on a project basis...
So I'd assume that, just like with blockchain & VR/AR, the biggest value will be in very specific use cases in very specific verticals, as opposed to the "spray & pray" approach, offering an "AI as a commodity".
Or is this really just a EU FOMO play, i.e. US is doing it so Europe can't stay behind? (Probably hoping for EU funds/grants via that "Cedric O" guy.)