news

in the past two days, ai startup ssi （ Safe Superintelligence ）raised $1 billionthe news has attracted considerable attention.

in fact, in the ai ​​industry, 1 billion in financing is not a big deal.

but ssi is particularly noteworthy, mainly because this companyonly 3 months old，the team is only about 10 peopleso far,no productsyou can't even figure out what they're going to do.

its ceo daniel gross（ Daniel Gross ）they just told everyone that they intend to focus on "directly building secure superintelligence" . . .

despite this situation, they were able to raise $1 billion, which is probably a money-making legend in the minds of many entrepreneurs.

but to be honest, it’s really hard for everyone to learn ssi’s way of making money.

because ssi can get so much money,most likely ilya sutskov（ Ilya Sutskever ）the golden signboard。

regarding ilya, everyone's memory probably still lingers on last year's openai palace farce, when the openai board of directors fired founder and ceo altman, and chairman greg was also fired.

the final outcome was against tiangang, and ultraman staged a comeback as a king.

ilya, who was considered by many to be the mastermind behind the incident, was repeatedly marginalized and officially resigned in may this year.

after leaving, elijah did not stop for a second and directly brought in former y combinator (one of the most influential institutions in the technology startup industry)partner gross and former openai engineer daniel levy（ Daniel Levy ）founded ssi together.

from left to right: gross, ilya, levi

maybe everyone is not clear about how powerful ilya is?

in the past 20 years, ilya's published papers have been cited more than 427,000 times.is the most cited computer scientistone。

his work has changed the course of research in computer vision, machine translation, and natural language processing.。

as for elijah’s wonderful story, in fact, it was as uncontrollable as a cool novel as early as when he was learning from his master.

ilya was born in the soviet union, but from the age of 5 he went to jerusalem and later moved to canada.

it was also at the university of toronto in canada that he met and became a disciple of the father of deep learning——geoffrey hinton（ Geoffrey Hinton ）。

mr. hinton is even more top-notch;it can be called the only ancestor of the current large model。

before generative big models became popular, the ai ​​community had been locked in heated discussions about the direction of their research. the neural networks and deep learning that hinton was primarily researching were once abandoned by everyone, leaving almost only this one field left in the world.

at the most exaggerated time, even hinton's then mentor, professor higgins, switched to other areas of artificial intelligence. higgins even persuaded hinton to do the same and start over.

in the end, xinton was the only one defending the high ground. after developing to the sixth level of equipment, he pushed back the high ground of other ai routes.

so, to this day,you can almost see that the top teams of generative large models are basically all members of xinmen。

and elijah himself is the chief disciple of xinmen.

according to hinton's exclusive interview with sana, as early as 2003, elijah, who was still a sophomore, ran to the door of hinton's office and banged on it after washing dishes for a summer.

after meeting the big boss hinton, elijah was still arrogant: "i've been frying french fries for the whole summer, and i'm really tired of it. i might as well come to your lab to learn some deep learning for fun."

hinton was also confused at the time: "young man, you should make an appointment first, and then we can talk."

ilya: "how about we talk now?"

such a small story is enough to illustrate elijah's straightforward personality. he is even a little bit difficult to get along with others. this is why the dispute over openai's route became so heated.

anyway, hinton didn't get along with him at that time, so he just threw a paper to elijah and told him to go back and read it until he understood it.

a week later, elijah came to find hinton: i do not understand ” 。

hinton was a little speechless at the time: he recommended himself so loudly, and i thought he was quite capable, but it turns out that he couldn't even understand the basic derivation of the chain rule?

as a result, ilya said, i understand this, but i don’t understand why a better optimizer is not used to handle the gradient?

hinton's eyes suddenly lit up. you know, this was a problem that hinton's team had spent several years solving, but elijah only took a week to get to the heart of the problem.

similar stories have been repeated in elijah's learning career. even hinton said:(ilia)the original intuition about things is always very good.

this may be why geniuses are geniuses, just like yilin when he was a child.

maybe in ilya's eyes, deep learning also has a kind of symmetrical beauty.

in 2012, ilya participated in the imagenet competition held by stanford university and won the championship with the alexnet system he created.

they then published a paper on alexnet, which became one of the most cited papers in the field of ai.

we can sayelijah's work laid the foundation for most subsequent ai scholars.

later, elijah also worked with his mentor hinton and alex kriszewski(another top deep learning expert)a company called dnnresearch was established.

dnn is even more special than ssi today. this company did not have any products at the time, nor did it have any plans to research any products. to put it bluntly, their company's products were the brains of the three of them.

they decided to auction dnn, that is, themselves, to the world.baidu, google, microsoft and deepmind are competing.

when the bidding reached $44 million, they paused the auction and took the talent directly to google.

in their view, the right platform is more important than a higher price.

just like that, baidu regretfully missed out on the team that would change the world of ai in the future.

during his years at google, elijah did not contract big company disease.

he worked as a research scientist at google brain and realized the use of neural networks for translation, beating all translators at the time.provides a major upgrade to google translate.

in 2015, ilya met ultraman and greg, and immediately felt that he had found a like-minded team.

soon elijah rejected google's million-dollar annual salary and became the co-founder and chief technologist of openai.

everyone has heard the rest of the story, which is how, with elijah's key contributions, openai climbed step by step to the throne of ai, and then parted ways.

today, elijah’s position in the industry is unquestionable, and the heyday of large models brought about by deep learning has also made countless people look forward to the true ai era.

but no matter whether it is elijah or hinton, no matter how they call for ai safety issues on various platforms and media, it can no longer stop everyone's desire for technological upgrades.

just like after ilya left, openai immediately removed the "long-term risk" team. this team was responsible for developing technology, formulating frameworks, and improving security measures to ensure that openai's ai products are safe enough.

so, at this time, it is obviously a good thing that elijah established ssi, hoping to shift everyone's enthusiasm for agi to ai safety.

if possible, shichao hopes that elijah can be a wonder boy again.create an ai that is both safe and powerful enough to fight openai to the end。

we can also see whether elijah takes the story line in which he is the man destined to take down the "evil sect" alone, or takes the story line in which he loses the help of the journey to the west and the magic is no longer there.