YEET -  Innovation & Technology

By YEET MAGAZINE | Posted on September 16, 2021 at 7:32 am

How to make trillions of dollars?

Create an Artificial Intelligence business.

DeepMind had a relatively simple task.

In 2013, DeepMind created an AI to play Atari's arcade game: Breakout . AI was using data and pattern recognition to improve its capabilities.

After just 600 games, the AI ​​was playing like an expert.

It was a real technical feat.

Vicarious, another AI company, has taken on a much more complex task.

They used DeepMind's AI and made him play a weathered version of the game. The game was barely different from the original but the AI ​​failed to adapt and had a beginner's level. .

The Vicarious co-founder said the following:

“We humans are not only programmed to understand shapes and movements but we can also create models from the things we see. "

As humans we can very quickly adapt to new situations, Vicarious and a company focused on the creation of a General Artificial Intelligence, their goal is to:

“Combine what we currently know about Artificial Intelligence with what we know about Neuroscience. "

Indeed, they want to create a machine that thinks like a human.

They took the idea to investors and everyone invested. Vicarious' investor portfolio includes Jeff Bezos, Mark Zuckerberg, Samsung and many more.

The idea would be to link AI to traditional robots that facilitate work: factory work, catering, problem solving etc ... The possibilities and concrete applications are endless.

Remember the name of Vicarious because they work in secrecy but they will hold the market in all areas, they will write the future. At some point, for better or for worse, these robots will get better than humans.

A better employee, a better baker, a better painter, a better… you.

Comments

How do you explain the differences between IBM's Watson and Google's DeepMind to non-specialists? from Google to non-specialists? from Google to non-specialists? from Google to non-specialists?

How does it feel to work at Google DeepMind ? ? ? ?

How can we see that Google is losing DeepMind ? ? ? ?

What is the storage size of a mainframe using AI like Google's Deepmind ? from Google? from Google? from Google?

How is Google Brain different from DeepMind ? ? ? ?

What do you think of the new artificial intelligence (AI) from DeepMind , a subsidiary of the parent company of Google Alphabet, which can detect breast cancer "as well" as a doctor?, a subsidiary of the parent company of Google Alphabet, can detect breast cancer "as well" as a doctor?, a subsidiary of the parent company of Google Alphabet, can detect breast cancer "as well" as a doctor?, a subsidiary of the parent company of Google Alphabet, can detect breast cancer "as well" as a doctor?

How could Google DeepMind employees spend their time better?

How did DeepMind start ? ? ? ?

What is the impact of the achievement of DeepMind in December 2021 in chess, on the future of AI? in December 2021 in chess, on the future of AI? in December 2021 in chess, on the future of AI? in December 2021 in chess, on the future of AI?

What could be the implications of artificial intelligence in the game Starcraft 2 (AlphaStar by Google's Project Deepmind ) for technological advancements in military tactics? Google) for technological advances in military tactics? Google) for technological advances in military tactics?

Louis Philippe Doyon-Lessard

· To follow

software developer Updated December 5, 2021

The arrival of artificial intelligence capable of operating in real time in video games may seem trivial, but the techniques that are developed for this will extend into other areas such as decision-making for the allocation of resources.

We notice that artificial intelligences are becoming more and more general. A few years ago, they could compete with humans only in extremely limited situations. But now, AlphaStar's performance in a complex game (Starcraft 2) shows us that AIs have decision-making abilities in situations that require much more general skills.

Some of these decision-making abilities will become greater in AI than in humans. We can imagine that armies will use AI to improve their decision-making, even in real time. According to several sources the US military has already started using AI and promises to use it ethically. ( The US military will use artificial intelligence "ethically"

) If they use AI, it is certainly who sees a substantial gain in their operation.

But back to Alphastar, there are a few differences with previous implementations that have caused it to play at a champion level:

Alphastar plays against itself even in practice;

Alphastar plays against an allied opponent (here the opponent wants Alphastar to improve, rather than trying to be victorious) also in practice

The first change is already used during war games. This is the simulation of war.

As far as I know, the second tactic, where some of the players don't aim to win, is not yet in use among the military. It might be interesting during war simulations to have units that want their opponent to improve in order to win real combat after training. This is something army leaders could learn from Alphastar by changing the way they play their war games. War games might include an opponent who looks like a trainer in order to recall past mistakes and avoid repeating them.

Also, on several occasions, artificial intelligences have surprised humans after training against themselves for a very long time, their random research allowing them to discover new tactics and strategies that were not yet known to the world. man.

Suppose that an army possesses an ultra-realistic simulation of the fights (much better than counter strike), the artificial intelligences could practice there and discover tactics for the moment unknown to our adversaries. It is by studying artificial intelligence that humans will learn. There are just too many possibilities in the real world to test everything, it's possible to test a lot more in simulated environments because you can run millions of them in parallel. We can therefore imagine that it is possible for the military to learn from artificial intelligence and transfer this knowledge to the real world.

162  views

What will Google DeepMind do with its $ 1 million AlphaGo prize? Is he doing the most efficient thing he could do with his prize money?with its $ 1 million prize from AlphaGo? Is he doing the most efficient thing he could do with his prize money?with its $ 1 million prize from AlphaGo? Is he doing the most efficient thing he could do with his prize money?

What are the companies that nobody knows but that will dominate the market in the future?

Ilkhom Mirzakirov

· To follow

Visual production in Act of Expression ( 2021 - today ) translator · September 21, 2021

Aled jenkins

B.Sc

This response may not be an accurate translation of Aled jenkins

in YEET in English  :What are the companies that no one knows about but could dominate future markets?

How to make trillions of dollars?

Create an Artificial Intelligence business.

DeepMind had a relatively simple task. had a relatively simple task. had a relatively simple task.

In 2021, DeepMind created an AI to play Atari's arcade game: Breakout . AI was using data and pattern recognition to improve its capabilities.. AI was using data and pattern recognition to improve its capabilities.. AI was using data and pattern recognition to improve its capabilities.

After just 600 games, the AI ​​was playing like an expert.

It was a real technical feat.

Vicarious, another AI company, has taken on a much more complex task.

They used DeepMind's AI and made him play a weathered version of the game. The game was barely different from the original but the AI ​​failed to adapt and had a beginner's level. . and had them played a weathered version of the game. The game was hardly different from the original but the AI ​​failed to adapt and had a beginner's level. and had them played a weathered version of the game. The game was hardly different from the original but the AI ​​failed to adapt and had a beginner's level.

The Vicarious co-founder said the following:

“We humans are not only programmed to understand shapes and movements but we can also create models from the things we see. "

As humans we can very quickly adapt to new situations, Vicarious and a company focused on the creation of a General Artificial Intelligence, their goal is to:

“Combine what we currently know about Artificial Intelligence with what we know about Neuroscience. "

Indeed, they want to create a machine that thinks like a human.

They took the idea to investors and everyone invested. Vicarious' investor portfolio includes Jeff Bezos, Mark Zuckerberg, Samsung and many more.

The idea would be to link AI to traditional robots that facilitate work: factory work, catering, problem solving etc ... The possibilities and concrete applications are endless.

Remember the name of Vicarious because they work in secrecy but they will hold the market in all areas, they will write the future. At some point, for better or for worse, these robots will get better than humans.

A better employee, a better baker, a better painter, a better… you.

6.8k  viewsShow 76  positive votesShow shares

76

9

3

What are the leading companies in the field of artificial intelligence? Why ?

Nicolas miné

· To follow

October 30, 2021

Hello,

It all depends on the domain but in general I would say that DeepMind

and OpenAI

are the 2 leaders in the field of AI.

The 2 are very innovative DeepMind is at the origin of the Deep-Q Networks and are the creators of AlphaGo. OpenAi has a lot of innovations including OpenAI Five.is at the origin of the Deep-Q Networks and are the creators of AlphaGo. OpenAi has a lot of innovations including OpenAI Five.

If we look at the publications level, it is also impressive the number of publications ( DeepMind published 72 publications in various subjects in 2021 for example: published 72 publications in various subjects in 2021 for example: Publications |

DeepMind

)

590  viewsShow 2  positive votes

2

Why does France claim that it has nuclear weapons when no one sees any kind of test from it, should the UN still believe this claim?

Pierre Leclercq

· To follow

Location: FranceFebruary 5, 2021

Nobody sees nuclear testing from France?

What not to read ...

After all the “patacaisse” (1) that we were given in 1995 when Jacques Chirac ordered to carry out the final French nuclear tests ...

NUCLEAR TEST: HARO SUR CHIRAC, RIOTS IN TAHITI

Outcry against the resumption of French nuclear tests

Me, during that time, I sleep soundly ...

The Chichi was brave to face the entire Earth so that French scientists could accumulate enough real-life experiences to refine computer models allowing nuclear reactions to be simulated with extreme precision.

And after that, shoo!

France has signed the CTBT (Treaty for the Complete Ban of Nuclear Tests).

Comprehensive Nuclear-Test-Ban Treaty - Wikipedia

We don't do real nuclear tests anymore, because we don't need them anymore.

We did not wait for DeepMind to do advanced simulations. to do advanced simulations.

In the meantime, we still have our nuclear submarines capable of plying the seas of the globe for centuries, carrying warheads ready to be used, in all invulnerability, if President Macron so decides.

(1)

Patacaisse - wiktionary

Pataques - wiktionary

2.4k  viewsShow 39  positive votesShow shares

Should companies serious about artificial intelligence use TensorFlow or should they create their own?

Adrien Lucas Ecoffet

· To follow

Studied Computer Science at Georgia Institute of Technology ( Diploma obtained in

A business serious about AI wants to spend its time doing AI, not optimizing low-level code in CUDA and C ++. Even the most serious AI companies, those that do basic research, use Tensorflow and other frameworks.

The frameworks that these different companies and research groups have used recently according to their githubs:

Google Brain: Caffe and Tensorflow

DeepMind: Torch and Tensorflow

OpenAI: Tensorflow

Allen Institute for Artificial Intelligence (AI2): PyTorch

Montreal Institute for Learning Algorithms (MILA): Theano

etc.

1.9k  viewsShow 11  positive votes

11

If Google employees were swapped with Microsoft employees, what would happen?

Tristan roth

I beg my friends to launch their projects translator · December 27, 2021

David Seidman

10 years as a Microsoft Program Manager, now at Salesforce

This response may not be an accurate translation of David Seidman

in YEET in English  :If Google employees were swapped with Microsoft employees, what would happen?

It's such a big “what if” that I couldn't pass it up, having worked for both companies.

In the beginning....

Google employees wake up at Microsoft : Hey, none of these tools work together! And our code bases are all fragmented and of different styles. It'll take forever to fix! Let's start right away. Wait, where are the snacks?

Microsoft employees wake up at Google : Wow, everything works here! I'm going to freeze a bit so that our team can achieve our goals. <repeated throughout the company> Wait, where is my office?

Few weeks later....

At Microsoft: Windows is now open source. Office has DeepMind Clippy who writes your documents for you. It doesn't really work, but it's a good idea. XBox games work on Android, if your phone is the right model (it isn't).

At Google : GMail has been renamed Google Mail Inbox for Enterprise Cloud Live. Google

continues to decline, but they figured out how to put an ad for Google Cloud in its place, and 7 Fortune 100 companies have done cloud deals, possibly as a result.

Months later ....

At Microsoft : DeepMind Clippy is now obsolete. DeepMind Messenger, Microsoft Chat, and an app named CallMe have all been launched and are mutually incompatible. Several googlers starved to death due to the lack of snacks. The autopsy was able to conclusively identify the root cause and a program was created to deliver snacks by unicycle when a Googler appears about to starve.

At Google : Android has been renamed Windows for Phones. It's 20% slower, but nothing else has changed (and never will). All the GBikes have been locked in the building that houses the most influential vice president. Absolutely no decision has been made regarding the products. Google Cloud now has 60% of the Fortune 500 as customers.

Soon after ....

Microsoft changes its name to Google. Google changes its name to Microsoft. And the balance is restored in the universe.

343  viewsShow 2  positive votesShow shares

2

1

What are the best AI research centers in Europe?

Haris sahovic

· To follow

Heading, I would say the Facebook research center in Paris [ 1 ] , headed by Yann Lecun [ 2 ] , one of the most important Deep Leraning specialists in the world, or the DeepMind [ 3 ] labs ( now a owned by Google), in London , and in particular at the origin of AlphaGo [ 4 ] .

The question undoubtedly deserves further research, however.

Footnotes

[ 1 ] Facebook AI Research (FAIR)

[ 2 ] Yann Le Cun - Wikipedia

[ 3 ] About Us | DeepMind

[ 4 ] AlphaGo - Wikipedia

387  viewsShow 1  positive vote

1

What is Deep Blue (Artificial Intelligence)?

Jean-Francois Dufour-Poirier

· To follow

Benevolent Soul ( 2021 - present ) October 25, 2021

The answers given are correct, but I want to add my note; that is to say that we must not confuse Deep Blue and Deep Mind.

Deep Mind is one of the best performing systems out there.

AlphaGo Zero: Google's AI

DeepMind

becomes self-taught

DeepMind

, Google's AI, declares war on StarCraft II

Interesting subject, recently Blizzard updated the AI ​​of some of these games, including heroes of the storm, and as a matter of fact, players are complaining that it's made too difficult.

In short, subject to follow with interest.

591  viewsShow 2  positive votesResponse requested by Achraf Karoui

Deepmind

Deepmind

To follow

Besides Coursera, what are the best online AI training sites?

A-Alexandra Boca

·

To follow

Artificial intelligence consultant at IBM (company) ( 2021 - present ) November 10, 2021

I like the DeepNLP and Deep learning ( deepmind ) courses which can be found on YouTube.

65  viewsShow 1  positive voteResponse requested by Younes Khadraoui

1

Which company is the most advanced in the field of artificial intelligence?

Nicolas miné

· To follow

February 10, 2021

Hello,

It depends on the field but in terms of R&D we find in top 3:

DeepMind ( Publications | DeepMind

)

OpenAI ( OpenAI Research

) -

Facebook ( Publications

)

For the first 2, it's “pure” AI with reinforcement learning. Facebook is more of machine learning.

55  views

How could Google choose to become a foundation in the service of humanity?

Nicolas caumette

· To follow

Curious about the human adventure and its improbable future February 5, 2021

At the moment, Google is an American commercial enterprise.

However, I see several things:

When I use Google's tools , I have the feeling that the company's approach is to make my life easier. Of course, it benefits from it through advertising placements, but I find that I first feel a satisfaction that I do not have with others, for example not with Microsoft. The repeated use of the search engine, like that of Maps is a very satisfying experience and I measure the effectiveness when I try to search differently.

Google bought Deepmind in early 2021 and yet when you go to the site http://Depmind.com

Google does not appear directly. First, Deepmind presents its mission: to use artificial intelligence in the service of humanity, in the fields of science, health, energy and more.

It is only by opening the Applications tab that we see Deepmind for Google appear . So Deepmind is primarily a research company, which publishes many articles on the latest advances in deep learning techniques ( deep learning ). And when Deepmind joined Google, they asked for the creation of an ethics committee, in order to ensure that their new owner would not impose on them behavior contrary to their mission of general interest.

In June 2021, Google announced that it was giving up the renewal of an artificial intelligence program for the Pentagon, following a petition from 4,000 of its employees and the threat of resignation of a dozen of them.

Since the summer of 2021, Google has maintained its desire to return to China with a version of its engine that satisfies censorship but faces an internal challenge from employees.

The CEO's argument is not to want to deprive 20% of the world's population of the benefits of technology and even if the commercial dimension is certainly at the center of the process, we must not exclude the soft power of Google. By maintaining its visibility in the Middle Kingdom, Google would keep a door open to Asia.

When we see that video game companies like Epic Games have reversed the strategy that prevailed among masters in the field like Sony, by offering Fortnite on all platforms and with the same level of access for all, we must ask ourselves the question. question of the interest of a more open approach, at the service of the greatest number.

In this context, it seems to me that the distance between the for-profit enterprise and a foundation that serves everyone is diminishing.

A foundation could ultimately be the best way to be chosen by the greatest number in the face of alternative proposals less committed to the general interest, or even directed against it.

This is just a hunch on my part, and your reactions are welcome!

What are the promising Artificial Intelligence companies to invest in?

Philippe Louis Delacan

Master in Machine Learning  and Imaging , École des Ponts ParisTech (ENPC) ( Diploma obtained in 2021 ) May 8, 2021

To make profit: Google Deepmind , Google Brain, Uber, Amazon, (Facebook?), Microsoft, Criteo, Boston Dynamics, Tencent

To benefit the planet and people: to look for

Who are the most recognized / well-known data scientists in this field?

Hello,

I don't know a lot of them but I would say my ranking would be:

Goeffrey Hinton and for good reason, he has a Wikipedia page

. He is also the inventor of the Networks capsule in 2021 which is not marked on WikiPedia

Yann LeCun who also has a WikiPedia

and one of the inventor of Deep Learning

Andrew Ng also with a WikiPedia page

a little less complete but still

Siraj Raval Youtubeur and

director of Home - School of AI

Aurélien Géron - former Google employee in the field and author of many books ( Amazon

)

A little off topic because he no longer works directly in the field, there is Demis Hessabis (CEO of Deepmind).

If you want other people to follow on twitter (all the previous ones have an account, I can also quote:

@OriolVinyalsML

@OpenAI

@DeepMindAI

@sirajraval

@karoly_zsolnai

@xsteenbrugge

I use it to follow the news in addition to the more “classic” accounts

How is AlphaFold revolutionary?

Luc Masure

· To follow

Formerly Retired Higher Education at HEFF Brussels ( ) December 1, 2021

From what I know, it is especially Deepmind and its artificial intelligence project based on neural networks that is revolutionary. Alphafold seems to me to be an application of this philosophy and does not seem to me to be its own, revolutionary even if its algorithm is certainly very intelligent (but not artificial).

74  viewsShow 1  positive vote

1

How does Elon Musk work on multiple missions simultaneously?

Tristan roth

· To follow

I beg my friends to launch their projects translator · April 18, 2021

Asim Qureshi

CEO Jibble | Meet the new standard in employee time tracking

This response may not be an accurate translation of Asim Qureshi

in YEET in English  :How is Elon Musk simultaneously working on multiple missions?

Obviously he needs to delegate, and delegate very well. But he knows he won't be able to delegate to an Elon, so he has to get involved, and since he only has a percentage of time on each business, he has to understand enough to make important decisions.

Many Elon employees know more about their area of ​​expertise than he does, but they're not him - he has his background and his own money at stake.

Elon's multiple projects greatly benefit his businesses through cross-learning. Being deeply involved in multiple companies, he continuously feeds cross-learning between companies at the highest level. This is what I feel almost daily with the companies that I have co-founded.

But there are other advantages. He is undoubtedly more connected by managing several companies. So when it comes to fundraising, for example, he might just have completed a round of funding for Tesla, and then, three months later, he brings in the same investment bankers for SpaceX who will have. deal with investors already comfortable with Elon. And his own brand, which comes from several companies, brings a lot to the table.

Elon did not rewrite rules by running multiple companies simultaneously. In fact, what he does has been proven to work. Google is no longer just a search engine - Android, GV, Google Capital, Google DeepMind , Google X, and a host of other diametrically tangential companies are all headed by Sergey and Larry. Go to Korea and you will find that Samsung and other chaebols have been doing the same thing for years. And Richard Branson did it with Virgin Group.