> Run:ai has developed an orchestration and virtualization software layer tailored to the unique needs of AI workloads running on GPUs and similar chipsets. Run:ai’s Kubernetes-based container platform for AI clouds efficiently pools and shares GPUs by automatically assigning the necessary amount of computing power – from fractions of GPUs, to multiple GPUs, to multiple nodes of GPUs.
Basically Nvidia coming out with managed AI services on DGX cloud
"NVIDIA is partnering with leading cloud service providers to host DGX Cloud infrastructure, starting with Oracle Cloud Infrastructure (OCI). Its OCI Supercluster provides a purpose-built RDMA network, bare-metal compute and high-performance local and block storage that can scale to superclusters of over 32,000 GPUs.
Microsoft Azure is expected to begin hosting DGX Cloud next quarter, and the service will soon expand to Google Cloud and more."
DGX cloud is nvidia's cloud offering (you can think of it like a snowflake/databricks in that it runs on top of other clouds), with allegedly the reference design architectures vs someone just slapping some H100s in a server rack.
I suppose we're now entering the era where Nvidia is so big and wealthy that they will join the other kings (e.g. FAANG+) in being acquisition machines. The contemporary regulatory environment may make this an interesting thing to watch. I suspect that they'll (by that I mean most tech companies) have to get these acquisitions done earlier while the companies are even smaller if they are to get past regulators.
We are well into that era already. Nvidia tried to buy ARM from softbank but was blocked by regulators a few years ago.
It's amazing that a company known for video game graphics cards is now a $2 trillion company. Nvidia's market cap ( over 2 trillion ) is the 2nd biggest of the FAANG+. It's bigger than Google. It's bigger than Amazon. Only Apple has a bigger market cap. And most of nvidia's market cap gains has been in the past year. It's unique for being the only predominantly hardware company of the group.
Both Apple and Nvidia use TSMC for chip manufacturing. Actually Apple might do a lot more of the hardware themselves, and more varied. Nvidia is just a design house, chips at TSMC and the surrounding electronics by many partners like Asus or Gigabyte.
It makes sense from the perspective if they are buying with equity. Their stock being at such heights its like getting a huge discount on your purchases and it barely goes on the balance sheet. The question always remains if the purchase was worth it - always tbd.
Also Netflix had greater ambitions back then which would be proven to be a red herring when they first spun out Roku and secondly focused on becoming “HBO before HBO can become us”. Reed probably didn’t have the chops for it, but one can imagine what Netflix could have actually become if it tried to build platforms.
Nvidia's footprint is more specific and obtuse than the other FAANG+s, so I expect they'll have an easier time with acquisition regulatory approval on the edges of their business model.
If they wanted to acquire an ML chip company, sure, but I/P/SaaS value adds will probably sail through.
They tried to buy ARM, and usa, eu, gb and china both said no. Lol poor nvidia. (I was so hyped for them, nvidia creating an arm desktop platform, and samsung/qualcomm/mediatek going against wintel, but that's not gonna happen now.)
i doubt NVIDIA main business have lots of differences, and this acquisition make sense both as add on for their product ecosystems (lock in) or too trying to pivot too cloud base solution.
This announcement doesn't mention a price, but the scoop by Calcalist last week says "many hundreds of millions of dollars and could even reach $1 billion." Good for them. (https://www.calcalistech.com/ctechnews/article/rkmbe9e0p)
You can use kubernetes to manage bare metal machines, but it probably doesn't make much difference: once the bits are on the GPU, I doubt a properly configured VM is getting in the way, let alone containerd.
But even if we set that aside you’ll get access to your data over a network connection because these are expensive nodes running batch jobs with finite disk space, not personal workstations.
Yea ofc. Nvidia has for mellanox infiniband, nvlink external, and even mellanox ethernet pci-to-pci, no need to involve the CPU. nvidia-docker has a few mods to support this too.
You can give GPU to VM and containers without any overhead these days. However, Nvidia always placed roadblocks when you aren't using their data center hardware in datacenters.
Honestly, I think it may be just a "marketing stunt" to feed the profitable hype for them, right now, in times of growing AI disillusion. Same with OpenAI and Moderna.
By reducing competition and diversity. Such acquisitions also often (but not always) results in the worsening of the products/services that the acquired company produces.
It's pretty difficult to come up with examples of acquisitions or mergers that resulted in something positive for the market and society at large.
M&A people specialize in making excuses as to why their merger will definitely be something positive that benefits the market and society at large. Of course, time usually proves them wrong and nobody is ever held accountable, but don't underestimate them, they are good at this.
If you google "Unit 8200" that is pretty much the reason. The Israeli military pushes their super-smart kids together to be computer scientists and hackers for 2 years - this is essentially the NSA of Israel. Behind every Israeli tech company is a small cadre of unit 8200 grads.
When they finish, a huge network of Unit 8200 alumni who are already rich have a shared experience with these kids, and are ready to invest in any idea these kids have. They have a tight-knit group of incredibly smart colleagues as well, from their time in the military.
> Behind every Israeli tech company is a small cadre of unit 8200 grads.
This is of course greatly exaggerated. I know plenty of Israeli startups that never had 8200 grads in them, mine included. I actually think that the role of a specific army unit in creating the environment that fosters these startups is really hyped. It's much more of a cultural thing than some random army unit.
Overall military spending can explain a lot as well. A lot of military spending goes to existing companies as well as startups to develop weapons and defence systems which drives innovation in high-tech technology.
Israel spends 5.3% of their GDP on military. That's more than almost any other country in the world and they have spent a lot for many decades.
I don't think the Israeli defense budget really goes into Israeli startups. The only related fields are drone startups and cyber startups that hardly sell domestically
I think it is more than 5 years on average. Here is the answer to prompt from Meta.ai below:
Unit 8200 is an elite intelligence unit of the Israel Defense Forces (IDF), and the length of service for new recruits can vary depending on their role and position.
Typically, new recruits in Unit 8200 serve for a minimum of 4-5 years, but some may serve longer depending on their specialization and the needs of the unit. Some roles may require even longer service commitments, up to 6-8 years or more.
It's worth noting that Unit 820psilon is considered one of the most prestigious and competitive units in the IDF, and service there is highly sought after by many young Israelis. The unit is responsible for military intelligence gathering, signal intelligence, and cybersecurity, among other things.
Even if the startup's founder is from 8200, the rest are not.
I knew one of the execs and the person is very smart.
It's just the culture that promotes intelligence. At least in Canada/US there is less emphasis on intelligence and you can see it on the VP level that they are more business-oriented and not necessarily the smartest person in the room, which is fine if you are in North America.
I'm eternally confused as to why other countries don't do this. The Manhattan Project resulted in so many great minds, purely due to the cross-pollination of all those geniuses in such a small place. You'd think any nation state would be gunning for the same advantage.
The US does kind of do this through its national labs. They just don't have the thriller-esque story of the Manhattan project to romanticize things and aren't wasting their intellect on maximizing doomscrolling.
Look at all the protests going on at places like Google. There's a segment of the population, especially in tech, that is irrationally anti-military. Somehow people think if we all held hands and sang Kumbaya, all the evil wars would stop, as if people like Vladimir Putin didn't exist.
There is also a work culture of both putting in as many hours as necessary, but also very goal oriented and understanding whats critical to deliver and what sint
There are obviously many more countries that make the average American or European tech worker seem lazy, but for some of them working a 60-80 hour week is about showing you're working rather than doing important work
From my experience its not strictly the hours put in per week driving Israeli startups, but actually delivering what customers are asking for. I've been in many calls where the executives promised a complex feature that hasn't even been planned yet be delivered very soon, and somehow the team is able to complete it with careful planning and prioritizing (and putting in the hours)
The executives dont shy away from the grind either, im now on the other side (US corporate) and Israeli startup CEO/CTOs are basically available for every concern I have 24/7, even when I dont expect them to answer they do
+1 In my previous stint, I had worked with Spot (https://spot.io/) as one of our vendors. Absolutely great product, amazing customer support and ability to take feature requests, or otherwise address our pain points quickly and effectively.
I have also experienced this. Given Israel is the opposite side of the world from California, mainly I expect if I ask a question in the afternoon there will be an answer ready in the morning. But surprisingly often, they answer immediately.
The harder problem is actually that the work week is Sun-Fri instead of Mon-Sun, which means there's days when people just aren't working. This is also sometimes a good thing though - Israelis will have put in a full day of work while everyone else was out for the weekend.
I suspect having a 10M population "only" + very bad commercial relations with their direct neighbours make Israeli startups think global from scratch. In Europe, most startups tend to think local-first, which hinders their scaling.
There's "Startup Nation" by Senor and Singer with a thesis on this.
My personal opinion is huge talent density, a just-get-it-done attitude, strong network effects from military service and a small number of universities, and a lot of startup experience across the board - if you're an entrepreneur in Tel Aviv, it's relatively easy to find people who've done it before multiple times, and they'll give you good advice.
Weizmann institute of science is the envy of the world in terms of their research output, and widely used as a model for new research institutes to aspire to.
This touches on some potentially sensitive topics but at least some of the success is explained by having a population with, on average, the second highest IQs.
Another part of it is definitely the benefits from having a strong high-trust culture that gives them access to social networks and a priority to provide for the group as a whole not just their own personal interests.
There is a very large percentage of population that doesn't participate in the economy and is largely uneducated. So it's even more shocking to see Israel so successful. And of course, you have the huge wave of immigration from the former USSR in the 90's.
Sadly I'm very pessimistic about the future of Israel and the demographics are against it.
I'm no economic historian and don't know where the relevant figures would be for Israel, but having skimmed some history of South Korea and Taiwan I get the impression that when a small country overperforms like this it usually involves a combination of generous domestic subsidies/incentives and a massive influx of foreign capital, which can come in numerous forms including technology transfer. Having a diaspora in more developed countries probably doesn't hurt a country's chances either (see e.g. Morris Chang).
The thing is, Jews tend to overperform even more outside of Israel. From Wikipedia:
Of the 965 individual recipients of the Nobel Prize and the Nobel Memorial Prize in Economic Sciences between 1901 and 2023,[1] at least 214 have been Jews or people with at least one Jewish parent, representing 22% of all recipients. For comparison, 65.4% of Nobel prize winners were either Christians or had a Christian background. Jews comprise only 0.2% of the world's population, meaning their share of winners is 110 times their proportion of the world's population.
That's the oppressor/oppressed narrative for any such difference. Hitler came to the same conclusion, and decided to do something about it. Various communists revolutionaries have used the same kind of thinking to get rid of the ruling classes in their countries after the revolution.
But if you consider it this way: The Nobel price was given mostly to white people (including Jews) until about 1970. That's a factor of 10 already.
Then you consider that Ashkenazi Jews have an average IQ of about 115 vs other Europeans somewhere south of 100 on average.
If the standard deviation is a constant 15, a population with a mean of 100 has about 0.4% of the population with an IQ >140. Increase the mean to 115, and that grows to about 5%. That's another factor of 10. And the gap widens as the threshold increases.
This seems to fit the 110x almost perfectly.
Disclaimer: I have no Jewish ancestry that I'm aware of.
Edit: Also, to my knowledge, the Nobel committee has NOT been dominated by Jews.
13% of the world population have an IQ higher than 117. That means that there is a billion non jew that has an IQ higher than the average Ashkenazi. Thus your arguments makes no sense. Even if every single jewish person had a high IQ, they will be less than 1% of the pool.
Well, there is the part about the selection pool, but there is also how a normal distribution behaves far out on the tail.
Most Nobel Laureates in the sciences are highly intelligent, so they exist on that tail.
Basically, if you sample from the top 13% from a normal distribution with (mean, std)=(100,15) you get fewer with 140+ than if you sample from one with (mean, std)=(115,15), and the higher the intelligence, the bigger the difference will be.
This follows from the squaring of the exponential term in the standard distribution.
While it's possible that the shape or standard deviations for various population doesn't match my assumptions, but if they do, the observed number of Nobel prizes given to Jews is very close to what one would expect.
“Start up nation” does a good job exploring this topic. In short, it’s about never accepting authority and challenging everything, the roots for that come from the culture and even the religion.
It's this plus being a small country. The entire population of Israel is roughly the same as the Bay Area, so everyone is 2-3 degrees away from each other. This makes it easier to raise capital and build a startup.
The success of Checkpoint and PANW kickstarted the scene.
> I’m suspecting high levels of education plus a culture that encourages entrepreneurship and risk-taking
There's also the small matter of $200bn in US military assistance since 1973, which involved significant knowledge transfer across many different scientific and engineering domains.
The govt is westernized and allow relatively more freedom of thought than other nations. So the situation is similar to America. Plus Israel is always been focused on tech because they know they need leverage to defend itself as a small country surrounded by adversaries
This may be controversial to some people here, but Israelis have high IQs genetically and from good parenting. When you put a bunch of determined high IQ people together they're going to outperform.
Mandatory military service with a military that has a high weight on technical warfare/superiority is an oft given reason.
Another is network effect. The same way that Silicon Valley has a lot of very successful startups. Once you have a few in a region, it creates a knock on effect.
It inspires people to follow suite. It means there’s more money being pushed around in that area. It means there’s more people looking at developments in that area. Often someone may be lower down in one successful startup and then leave to make their own with the experience of how to do it.
I haven't researched deeply but there are plenty of Israeli startup exits with executives coming from combat, non tech units. Usually elite combat units that are harder to get in to.
There is something to be said about spending 18-21 in the military, there is pretty much 0 partying during college (that is done during military vacations + the gap year between military and college). By the time you've finished your Bachelors degree you're 25-26 and ready to settle down. You hit your full time job life 6-8 years after high school. Its a very different mindset
I look at the 22 year olds moving to NYC after college for their first job, and there is such a huge difference in what they're looking for in their day to day lives. Their job is probably the last thing they are about
Through the Israeli military service there is a large pipeline of (technical acumen cultivated while serving) -> (business || academia). Especially in technology. A large part of this is comes from the fact that Israel strives to be self sufficient when it comes to military technology.
When it comes to deals, there is also, frankly, a large Jewish funding/investing community. The hard part of getting deals is who you know.
My assumption is all of the above, as well as well connected folks through world class educational institutions, and a generally practical environment.
Investment into their people. Proper support structures. Taking advantage of their “country pride”.
Same effect was seen in USA during the space race.
At some point USA just “gave up” and ceded all control to private corporations which don’t give a shit about the people and only care about the bottom line
If you surround yourself by friends who all do that one thing and are all driven to be good at that thing, you will be motivated to do the same thing and get good at it, and motivate other people in turn.
Jews have a higher proportion of very smart people than other ethnicities. It's something that is "complicated" to talk about (we don't like all the implications of such a possibility), but it's pretty much indisputable.
You dont need intelligence to have a startup that succeeds. I bet that within this thread, there are users who have ideas that can become mutltimillion companies. All it would take is someone who knows how to navigate the business space to set up funding and talk to the right people in the right way to make sales.
Success in startup space is more about motivation in terms of pushing harder than others, and support system that gives you a higher chance to make correct decisions.
> You dont need intelligence to have a startup that succeeds
Not necessarily genius-level but being smart I'm sure helps. You might not need it, in the sense that there certainly are counter-examples, but yeah, on the big numbers it's pretty obvious that it's an advantage.
Usually it's the far left that use the oppressor/oppressed way of thinking to demonize their enemies. But when it comes to Jews, Nazis and many other extreme right wingers do the same.
Hitler saw the Jews as an existential threat, not as objectively inferior to "Aryans". In fact, this is why he hated them so much above anyone else; he thought that they were the only "Race" that had the potential to out-compete his "Aryans".
> Hitler saw the Jews ... not as objectively inferior to "Aryans". ... he thought that they were the only "Race" that had the potential to out-compete his "Aryans"
We're veering terribly off topic, but I'm pretty sure this is not historically true.
Frankly, when you start attributing racial attributes to me - even positive ones - and telling me you got these ideas reading Mein Kampf, I get very, very wary.
I understand what you mean, and this is one of the many reasons we all feign great surprise at this news and others. However, an undesirable or uncomfortable reality doesn't make it any less real.
And, even if you take the popular position that objectively extraordinary Jewish achievements are all due to education- see for example here: [1] - the bottom line doesn't change- there are more higher achievers in the Jewish population.
Nitpick: note that I'm not attributing anything to you, that would surely be racist. Every individual is different and needs to be judged on his or her own merits. Also, the Mein Kampf guy is the other one :)
Interesting link. I specifically noticed this part:
And the argument could be made that this is nothing to do with genes but instead a response to handicaps placed on our people since time immemorial by those who want us to fail.
This is interesting, because it's precisely such handicaps that could explain a genetic component to Jewish performance. And the Ashkenazi in particular.
If a population is banned, for hundreds of years, from owning land and working as farmers, that creates a very significant selection pressure for finding other ways to survive. Those not able to do so either died off or converted to Christianity.
And it just happened to be that most other livelihoods available to them required more mental capacity than farming.
Also, as you say, this is at the population level. It's perfectly possible for Jews to have average or below intelligence. It's simply a lot more likely for an Ashkenazi Jew to have an IQ of 160 than for other Europeans (or other Jewish groups).
> Run:ai has developed an orchestration and virtualization software layer tailored to the unique needs of AI workloads running on GPUs and similar chipsets. Run:ai’s Kubernetes-based container platform for AI clouds efficiently pools and shares GPUs by automatically assigning the necessary amount of computing power – from fractions of GPUs, to multiple GPUs, to multiple nodes of GPUs.
Basically Nvidia coming out with managed AI services on DGX cloud
Apparently the DGX cloud is hosted first on Oracle, moving to Azure and hosting on Google soon...
From 2023: https://nvidianews.nvidia.com/news/nvidia-launches-dgx-cloud...
"NVIDIA is partnering with leading cloud service providers to host DGX Cloud infrastructure, starting with Oracle Cloud Infrastructure (OCI). Its OCI Supercluster provides a purpose-built RDMA network, bare-metal compute and high-performance local and block storage that can scale to superclusters of over 32,000 GPUs.
Microsoft Azure is expected to begin hosting DGX Cloud next quarter, and the service will soon expand to Google Cloud and more."
There are also self-hosted DGX Grace Blackwell systems releasing later this year / next year.
DGX cloud is nvidia's cloud offering (you can think of it like a snowflake/databricks in that it runs on top of other clouds), with allegedly the reference design architectures vs someone just slapping some H100s in a server rack.
Run ai is not a mature product at this point though.
Nvidia definitely overpaid an order of mangitude here, this setup's a month ish work for any semi-skilled SWE who is familiar with RDMA, SR-IOV etc.
I suppose we're now entering the era where Nvidia is so big and wealthy that they will join the other kings (e.g. FAANG+) in being acquisition machines. The contemporary regulatory environment may make this an interesting thing to watch. I suspect that they'll (by that I mean most tech companies) have to get these acquisitions done earlier while the companies are even smaller if they are to get past regulators.
We are well into that era already. Nvidia tried to buy ARM from softbank but was blocked by regulators a few years ago.
It's amazing that a company known for video game graphics cards is now a $2 trillion company. Nvidia's market cap ( over 2 trillion ) is the 2nd biggest of the FAANG+. It's bigger than Google. It's bigger than Amazon. Only Apple has a bigger market cap. And most of nvidia's market cap gains has been in the past year. It's unique for being the only predominantly hardware company of the group.
I think Nvidia would argue their software is at least as valuable to their market position as Apple's is.
Microsoft is missing from your list. And if Nvidia is predominantly hardware, so is Apple.
Nvidia is a software company that happens to make their own chips. This has been their mantra for the last 15 years.
I thought they paid TSMC to make chips they design. Just like Apple.
Both Apple and Nvidia use TSMC for chip manufacturing. Actually Apple might do a lot more of the hardware themselves, and more varied. Nvidia is just a design house, chips at TSMC and the surrounding electronics by many partners like Asus or Gigabyte.
It makes sense from the perspective if they are buying with equity. Their stock being at such heights its like getting a huge discount on your purchases and it barely goes on the balance sheet. The question always remains if the purchase was worth it - always tbd.
Nvidia is the actual N in FAANG. Netflix is not a technological powerhouse compared to Nvidia
Netflix was included because they were a hypergrowth SV tech stock, paying crazy compensations, not because it was a tech powerhouse.
When FAANG was created, Netflix was the highest gainer in the S&P 500. FAANG was the highest growth stocks that year, and nothing more.
Also Netflix had greater ambitions back then which would be proven to be a red herring when they first spun out Roku and secondly focused on becoming “HBO before HBO can become us”. Reed probably didn’t have the chops for it, but one can imagine what Netflix could have actually become if it tried to build platforms.
Netflix is definitely a technological powerhouse, it just we, mortals, can't benefit from it.
I am expecting Groq acquisition or merger at some point. I just makes sense. Especially considering the financial position of Nvidia right now.
Nvidia's footprint is more specific and obtuse than the other FAANG+s, so I expect they'll have an easier time with acquisition regulatory approval on the edges of their business model.
If they wanted to acquire an ML chip company, sure, but I/P/SaaS value adds will probably sail through.
They tried to buy ARM, and usa, eu, gb and china both said no. Lol poor nvidia. (I was so hyped for them, nvidia creating an arm desktop platform, and samsung/qualcomm/mediatek going against wintel, but that's not gonna happen now.)
i doubt NVIDIA main business have lots of differences, and this acquisition make sense both as add on for their product ecosystems (lock in) or too trying to pivot too cloud base solution.
This announcement doesn't mention a price, but the scoop by Calcalist last week says "many hundreds of millions of dollars and could even reach $1 billion." Good for them. (https://www.calcalistech.com/ctechnews/article/rkmbe9e0p)
Techcrunch mentioned it was $700M
NVIDIA's vertically integrating their platform and making it the single place for your AI infra needs - pretty cool
I don't share your enthusiasm for monopolies
It's cool if you're Nvidia, work for them or own stocks. Not so much otherwise.
Reading Stratechery has got a lot of people talking like a buy-side equities analyst when discussing companies. Vendor lock-in is now a good thing. /s
Calcalist says the deal is valued at [$700 million](https://www.calcalistech.com/ctechnews/article/s1w2iniwa).
I was under the impresion that modern AI still runs mainly baremetal. Isn't pytorch parallel distribute not good enough?
You can use kubernetes to manage bare metal machines, but it probably doesn't make much difference: once the bits are on the GPU, I doubt a properly configured VM is getting in the way, let alone containerd.
the problem is multinode runs that communicate through the network
Multinode runs don’t communicate through the network in a DGX configuration. NVlink allows for RDMA over direct infiniband. No need for network here.
Infiniband is a network too…
But even if we set that aside you’ll get access to your data over a network connection because these are expensive nodes running batch jobs with finite disk space, not personal workstations.
Yea ofc. Nvidia has for mellanox infiniband, nvlink external, and even mellanox ethernet pci-to-pci, no need to involve the CPU. nvidia-docker has a few mods to support this too.
Yes, which is especially important for training. Getting good GPU interconnect can be really important for training large models.
There are a ton of tricks you can do to share hbm/flops amongst GPUs in a cluster.
> I was under the impresion that modern AI still runs mainly baremetal.
Generally yes, but you can do (PCI) pass-through of GPUs to VMs and containers and such, so it is not strictly necessary.
You can give GPU to VM and containers without any overhead these days. However, Nvidia always placed roadblocks when you aren't using their data center hardware in datacenters.
This is purely a scheduler, it doesn't have any any effect on the workload performance.
nVidia Multi-Instance GPUs has been around for >4 years.
I am so tired of all these acquisitions. My personal opinion is that it is ruining our economy.
yup. And most of these acquisitions end up just sitting on a shelf so the company can hoard the IP rights.
So nvidia can rest assured that nobody will touch them, or at least impede their progress significantly.
Honestly, I think it may be just a "marketing stunt" to feed the profitable hype for them, right now, in times of growing AI disillusion. Same with OpenAI and Moderna.
Care to explain how?
By reducing competition and diversity. Such acquisitions also often (but not always) results in the worsening of the products/services that the acquired company produces.
It's pretty difficult to come up with examples of acquisitions or mergers that resulted in something positive for the market and society at large.
M&A people specialize in making excuses as to why their merger will definitely be something positive that benefits the market and society at large. Of course, time usually proves them wrong and nobody is ever held accountable, but don't underestimate them, they are good at this.
Can someone explain how or why Israeli startups have historically punched above their weight?
For such a small country, they keep churning out tons of successful companies that get acquired by larger American and European counterparts.
I’m suspecting high levels of education plus a culture that encourages entrepreneurship and risk-taking, but there might be more to the story..
If you google "Unit 8200" that is pretty much the reason. The Israeli military pushes their super-smart kids together to be computer scientists and hackers for 2 years - this is essentially the NSA of Israel. Behind every Israeli tech company is a small cadre of unit 8200 grads.
When they finish, a huge network of Unit 8200 alumni who are already rich have a shared experience with these kids, and are ready to invest in any idea these kids have. They have a tight-knit group of incredibly smart colleagues as well, from their time in the military.
> Behind every Israeli tech company is a small cadre of unit 8200 grads.
This is of course greatly exaggerated. I know plenty of Israeli startups that never had 8200 grads in them, mine included. I actually think that the role of a specific army unit in creating the environment that fosters these startups is really hyped. It's much more of a cultural thing than some random army unit.
Overall military spending can explain a lot as well. A lot of military spending goes to existing companies as well as startups to develop weapons and defence systems which drives innovation in high-tech technology.
Israel spends 5.3% of their GDP on military. That's more than almost any other country in the world and they have spent a lot for many decades.
https://en.wikipedia.org/wiki/List_of_countries_with_highest...
I don't think the Israeli defense budget really goes into Israeli startups. The only related fields are drone startups and cyber startups that hardly sell domestically
[dead]
Plenty of communities have cultures of entrepreneurship and hard work, though.
I think it is more than 5 years on average. Here is the answer to prompt from Meta.ai below: Unit 8200 is an elite intelligence unit of the Israel Defense Forces (IDF), and the length of service for new recruits can vary depending on their role and position. Typically, new recruits in Unit 8200 serve for a minimum of 4-5 years, but some may serve longer depending on their specialization and the needs of the unit. Some roles may require even longer service commitments, up to 6-8 years or more. It's worth noting that Unit 820psilon is considered one of the most prestigious and competitive units in the IDF, and service there is highly sought after by many young Israelis. The unit is responsible for military intelligence gathering, signal intelligence, and cybersecurity, among other things.
Even if the startup's founder is from 8200, the rest are not. I knew one of the execs and the person is very smart. It's just the culture that promotes intelligence. At least in Canada/US there is less emphasis on intelligence and you can see it on the VP level that they are more business-oriented and not necessarily the smartest person in the room, which is fine if you are in North America.
I'm eternally confused as to why other countries don't do this. The Manhattan Project resulted in so many great minds, purely due to the cross-pollination of all those geniuses in such a small place. You'd think any nation state would be gunning for the same advantage.
Manhattan projects don't really work unless you feel that you are under existential threat.
Israelis perpetually do.
The US does kind of do this through its national labs. They just don't have the thriller-esque story of the Manhattan project to romanticize things and aren't wasting their intellect on maximizing doomscrolling.
Look at all the protests going on at places like Google. There's a segment of the population, especially in tech, that is irrationally anti-military. Somehow people think if we all held hands and sang Kumbaya, all the evil wars would stop, as if people like Vladimir Putin didn't exist.
Think the idea here is that the US supplies the weapons for those wars and we should stop doing that.
Which is a clueless idea, but it's an idea. The US refusing to supply weapons would have, if anything, encouraged Putin to be even more genocidal.
There is also a work culture of both putting in as many hours as necessary, but also very goal oriented and understanding whats critical to deliver and what sint
There are obviously many more countries that make the average American or European tech worker seem lazy, but for some of them working a 60-80 hour week is about showing you're working rather than doing important work
From my experience its not strictly the hours put in per week driving Israeli startups, but actually delivering what customers are asking for. I've been in many calls where the executives promised a complex feature that hasn't even been planned yet be delivered very soon, and somehow the team is able to complete it with careful planning and prioritizing (and putting in the hours)
The executives dont shy away from the grind either, im now on the other side (US corporate) and Israeli startup CEO/CTOs are basically available for every concern I have 24/7, even when I dont expect them to answer they do
+1 In my previous stint, I had worked with Spot (https://spot.io/) as one of our vendors. Absolutely great product, amazing customer support and ability to take feature requests, or otherwise address our pain points quickly and effectively.
> even when I dont expect them to answer they do
I have also experienced this. Given Israel is the opposite side of the world from California, mainly I expect if I ask a question in the afternoon there will be an answer ready in the morning. But surprisingly often, they answer immediately.
Many Israelis in tech worked shifted hours to line up with the USA better, because either their customers or coworkers are in the USA.
Some do, not all.
The harder problem is actually that the work week is Sun-Fri instead of Mon-Sun, which means there's days when people just aren't working. This is also sometimes a good thing though - Israelis will have put in a full day of work while everyone else was out for the weekend.
I would also assume it might be a good circadian fit for many, given the prevalence of "night owls" in tech
I suspect having a 10M population "only" + very bad commercial relations with their direct neighbours make Israeli startups think global from scratch. In Europe, most startups tend to think local-first, which hinders their scaling.
There's "Startup Nation" by Senor and Singer with a thesis on this.
My personal opinion is huge talent density, a just-get-it-done attitude, strong network effects from military service and a small number of universities, and a lot of startup experience across the board - if you're an entrepreneur in Tel Aviv, it's relatively easy to find people who've done it before multiple times, and they'll give you good advice.
It’s not only start ups, but also academia.
Weizmann institute of science is the envy of the world in terms of their research output, and widely used as a model for new research institutes to aspire to.
The same demographic is punching even more above their weight in the US.
Probably by a pretty large margin, too.
This touches on some potentially sensitive topics but at least some of the success is explained by having a population with, on average, the second highest IQs.
Another part of it is definitely the benefits from having a strong high-trust culture that gives them access to social networks and a priority to provide for the group as a whole not just their own personal interests.
There is a very large percentage of population that doesn't participate in the economy and is largely uneducated. So it's even more shocking to see Israel so successful. And of course, you have the huge wave of immigration from the former USSR in the 90's. Sadly I'm very pessimistic about the future of Israel and the demographics are against it.
The demographics are against the entire western world
however i wouldn't get too caught up with extrapolations, they are usually wrong (see population bomb)
If we look at IQ as an explanation, it's not only about the mean, but also about the variance, which is driven up by the Ashkenazi sub-population.
I would not be surprised if Israel has a significantly larger percentage of people with IQ>140 than any other nation.
I'm no economic historian and don't know where the relevant figures would be for Israel, but having skimmed some history of South Korea and Taiwan I get the impression that when a small country overperforms like this it usually involves a combination of generous domestic subsidies/incentives and a massive influx of foreign capital, which can come in numerous forms including technology transfer. Having a diaspora in more developed countries probably doesn't hurt a country's chances either (see e.g. Morris Chang).
The thing is, Jews tend to overperform even more outside of Israel. From Wikipedia:
Of the 965 individual recipients of the Nobel Prize and the Nobel Memorial Prize in Economic Sciences between 1901 and 2023,[1] at least 214 have been Jews or people with at least one Jewish parent, representing 22% of all recipients. For comparison, 65.4% of Nobel prize winners were either Christians or had a Christian background. Jews comprise only 0.2% of the world's population, meaning their share of winners is 110 times their proportion of the world's population.
[1] https://en.wikipedia.org/wiki/List_of_Jewish_Nobel_laureates
Unless they're somehow 110 times better than everyone else, this indicates some powerful nepotism.
Not true. The differences don't have to be that large.
If one person consistently runs 0.1s faster than everyone else, they can win 100x the races of the next best person without being 100x better.
Nobel prizes, like running races, are winner take all. The person in 100th place doesn't get any points proportional to how good their proposal is.
That's the oppressor/oppressed narrative for any such difference. Hitler came to the same conclusion, and decided to do something about it. Various communists revolutionaries have used the same kind of thinking to get rid of the ruling classes in their countries after the revolution.
But if you consider it this way: The Nobel price was given mostly to white people (including Jews) until about 1970. That's a factor of 10 already.
Then you consider that Ashkenazi Jews have an average IQ of about 115 vs other Europeans somewhere south of 100 on average.
If the standard deviation is a constant 15, a population with a mean of 100 has about 0.4% of the population with an IQ >140. Increase the mean to 115, and that grows to about 5%. That's another factor of 10. And the gap widens as the threshold increases.
This seems to fit the 110x almost perfectly.
Disclaimer: I have no Jewish ancestry that I'm aware of.
Edit: Also, to my knowledge, the Nobel committee has NOT been dominated by Jews.
13% of the world population have an IQ higher than 117. That means that there is a billion non jew that has an IQ higher than the average Ashkenazi. Thus your arguments makes no sense. Even if every single jewish person had a high IQ, they will be less than 1% of the pool.
Well, there is the part about the selection pool, but there is also how a normal distribution behaves far out on the tail.
Most Nobel Laureates in the sciences are highly intelligent, so they exist on that tail.
Basically, if you sample from the top 13% from a normal distribution with (mean, std)=(100,15) you get fewer with 140+ than if you sample from one with (mean, std)=(115,15), and the higher the intelligence, the bigger the difference will be.
This follows from the squaring of the exponential term in the standard distribution.
While it's possible that the shape or standard deviations for various population doesn't match my assumptions, but if they do, the observed number of Nobel prizes given to Jews is very close to what one would expect.
The selection pool is not the global population, as stated in the previous comment.
“Start up nation” does a good job exploring this topic. In short, it’s about never accepting authority and challenging everything, the roots for that come from the culture and even the religion.
> I’m suspecting high levels of education
It's this plus being a small country. The entire population of Israel is roughly the same as the Bay Area, so everyone is 2-3 degrees away from each other. This makes it easier to raise capital and build a startup.
The success of Checkpoint and PANW kickstarted the scene.
> I’m suspecting high levels of education plus a culture that encourages entrepreneurship and risk-taking
There's also the small matter of $200bn in US military assistance since 1973, which involved significant knowledge transfer across many different scientific and engineering domains.
The govt is westernized and allow relatively more freedom of thought than other nations. So the situation is similar to America. Plus Israel is always been focused on tech because they know they need leverage to defend itself as a small country surrounded by adversaries
This may be controversial to some people here, but Israelis have high IQs genetically and from good parenting. When you put a bunch of determined high IQ people together they're going to outperform.
Easy access to large VC companies in the US, networking effects.
Mandatory military service with a military that has a high weight on technical warfare/superiority is an oft given reason.
Another is network effect. The same way that Silicon Valley has a lot of very successful startups. Once you have a few in a region, it creates a knock on effect.
It inspires people to follow suite. It means there’s more money being pushed around in that area. It means there’s more people looking at developments in that area. Often someone may be lower down in one successful startup and then leave to make their own with the experience of how to do it.
I haven't researched deeply but there are plenty of Israeli startup exits with executives coming from combat, non tech units. Usually elite combat units that are harder to get in to.
There is something to be said about spending 18-21 in the military, there is pretty much 0 partying during college (that is done during military vacations + the gap year between military and college). By the time you've finished your Bachelors degree you're 25-26 and ready to settle down. You hit your full time job life 6-8 years after high school. Its a very different mindset
I look at the 22 year olds moving to NYC after college for their first job, and there is such a huge difference in what they're looking for in their day to day lives. Their job is probably the last thing they are about
Through the Israeli military service there is a large pipeline of (technical acumen cultivated while serving) -> (business || academia). Especially in technology. A large part of this is comes from the fact that Israel strives to be self sufficient when it comes to military technology.
When it comes to deals, there is also, frankly, a large Jewish funding/investing community. The hard part of getting deals is who you know.
My assumption is all of the above, as well as well connected folks through world class educational institutions, and a generally practical environment.
Nvidia has R&D operations in Israel for several years. Are any of the employees from Run:AI ex-NVIDIA?
Investment into their people. Proper support structures. Taking advantage of their “country pride”.
Same effect was seen in USA during the space race.
At some point USA just “gave up” and ceded all control to private corporations which don’t give a shit about the people and only care about the bottom line
When you've got the US fueling your war machine you tend to have a lot of money left over to plow into tech dominance.
I've worked for two, both inside and outside Israel, and both were very successful. The same thing happened both times.
1. Foreign Jewish VCs pumped money into it, gave the founders plenty of autonomy.
2. The internal network of the founders to other successful Jews was gigantic. Even if they're unimpressive themselves.
Essentially easy access to capital + domestic and international network that can facilitate success. It's a nationalized version of YC.
Israel is also small enough that almost everyone is 1-2 degrees away from a top tier vc
The boring answer is that its simple community.
If you surround yourself by friends who all do that one thing and are all driven to be good at that thing, you will be motivated to do the same thing and get good at it, and motivate other people in turn.
Silicon Valley used to be like this.
Probably a lot of money sloshing around the US-subsidized military-industrial complex?
> I’m suspecting high levels of education plus a culture that encourages entrepreneurship and risk-taking, but there might be more to the story..
I sincerely hope this is bait, because the answer is obvious.
Which is?
(I am genuinely asking)
Jews have a higher proportion of very smart people than other ethnicities. It's something that is "complicated" to talk about (we don't like all the implications of such a possibility), but it's pretty much indisputable.
You dont need intelligence to have a startup that succeeds. I bet that within this thread, there are users who have ideas that can become mutltimillion companies. All it would take is someone who knows how to navigate the business space to set up funding and talk to the right people in the right way to make sales.
Success in startup space is more about motivation in terms of pushing harder than others, and support system that gives you a higher chance to make correct decisions.
> You dont need intelligence to have a startup that succeeds
Not necessarily genius-level but being smart I'm sure helps. You might not need it, in the sense that there certainly are counter-examples, but yeah, on the big numbers it's pretty obvious that it's an advantage.
Hitler infamously condemned the IQ system for being "jewish". I don't think that sentiment has gone away.
Usually it's the far left that use the oppressor/oppressed way of thinking to demonize their enemies. But when it comes to Jews, Nazis and many other extreme right wingers do the same.
Hitler saw the Jews as an existential threat, not as objectively inferior to "Aryans". In fact, this is why he hated them so much above anyone else; he thought that they were the only "Race" that had the potential to out-compete his "Aryans".
> Hitler saw the Jews ... not as objectively inferior to "Aryans". ... he thought that they were the only "Race" that had the potential to out-compete his "Aryans"
We're veering terribly off topic, but I'm pretty sure this is not historically true.
That was my take when I read his section about Jews in Mein Kampf.
I'm Jewish. My grandparents fled the Holocaust and were part of the first generation of Israelis.
This isn't "indisputable," it's nonsense.
Frankly, when you start attributing racial attributes to me - even positive ones - and telling me you got these ideas reading Mein Kampf, I get very, very wary.
I understand what you mean, and this is one of the many reasons we all feign great surprise at this news and others. However, an undesirable or uncomfortable reality doesn't make it any less real.
And, even if you take the popular position that objectively extraordinary Jewish achievements are all due to education- see for example here: [1] - the bottom line doesn't change- there are more higher achievers in the Jewish population.
Nitpick: note that I'm not attributing anything to you, that would surely be racist. Every individual is different and needs to be judged on his or her own merits. Also, the Mein Kampf guy is the other one :)
[1] https://www.thejc.com/lets-talk/why-are-there-so-many-jewish...
Interesting link. I specifically noticed this part:
And the argument could be made that this is nothing to do with genes but instead a response to handicaps placed on our people since time immemorial by those who want us to fail.
This is interesting, because it's precisely such handicaps that could explain a genetic component to Jewish performance. And the Ashkenazi in particular.
If a population is banned, for hundreds of years, from owning land and working as farmers, that creates a very significant selection pressure for finding other ways to survive. Those not able to do so either died off or converted to Christianity.
And it just happened to be that most other livelihoods available to them required more mental capacity than farming.
Also, as you say, this is at the population level. It's perfectly possible for Jews to have average or below intelligence. It's simply a lot more likely for an Ashkenazi Jew to have an IQ of 160 than for other Europeans (or other Jewish groups).
[flagged]
This co facilitates efficient allocation of GPUs. I can almost say it’s like a hyper convergence service like Nutanix.