Things Are Moving Quickly: AI Chip Report Giving Some Insight

avatar
(Edited)

▶️ Watch on 3Speak


Most everyone has heard of Moore's Law. This is one of the key components that is responsible for the massive move forward in computing over the last 3 decades.

There is another processor pace that could make that look rather slow. AI processors are doubling at a rate, according to a report by Stamford University, that would put Moore's Law to shame.


▶️ 3Speak



0
0
0.000
19 comments
avatar

Dear @taskmaster4450

You've mentioned that processing power does double every 3 months. However does amount of data also increase that fast?

How much longer before increasing processing power won't change much. I remember that in the past we were lacking space for our data and hard drives were simply to small. Right now we can easily get hard drive which will cost very little and will store 500gb (or more) of our data (probably we will never even use it). Would I need newest disk which would be 50times "bigger"? not really.

I believe that with AI processing power we will also reach this point sooner or later. Wouldn't you agree?

Yours,
Piotr

0
0
0.000
avatar

The only thing that I can think to this is that: AI developments will remain a centralized projects for several years to a decade, focusing on specific area of usage like medical diagnosis, stem research, commerce, anything.

This will remain until hardware capabilities improve in terms of computational power.

The when 5g networks comes in and decentralization of AI developments have made its footing. The the BIG BANG OF AI TECHNOLOGY WILL HAPPEN.

I give it several decades more in the future.

0
0
0.000
avatar

Good morning @guruvaj

I just realized that I never actually thanked you for this amazing reply. Appreciate it a lot.

ps.
Would you mind if I take few minutes of your time? (I hope I'm not asking this question to often ;)

Anyway .... together with few core members of project.hope team - we'te trying to promote our recent publication: an article explaining economy behind our non-profit community project build on STEEM blockchain.

Perhaps I could ask you to spare few minutes and check it out and share your feedback with me.

I would absolutely appreciate it a lot. I read all comments and I drop solid upvote on each valuable one.

Link: https://steemit.com/hive-175254/@project.hope/3-ways-of-joining-our-efforts-project-hope-economy-explained

Yours, Piotr

0
0
0.000
avatar
(Edited)

Thanks for linking this video and your comment in my feed.

I believe we have enough data to utilize the processing power. There are data about anything we do stored somewhere in some data centers.

I don't think the size of data set would be our future bottleneck, instead the speed of reading/writing data. How fast is your disk transfer rate? How fast does the network transfer that data? Did we fully utilize all available nodes that can be used to do calculation in the same cluster? Does the program even know if there was an available node idling in the system? What is the most optimized scheduling algorithm? Questions can go on and on.

Well, at least this is my option.

In term of the price per processing power will go down (mentioned in the video), did we see cost of our data center go down? Or the price of our personal computer? This is a rich man's game.

Yes, we probably will see an affordable nVidia Tesla for personal use in a couple of years. However, by that time, big companies like Google, Facebook, or Research centers in MIT, Stanford have already moved to next bigger/better things. An individual/small entity just do not have the resource to fight against them.

I remembered a professor discussing about research in the small university where I work at. "When we are doing a research, and find out big companies like Google are doing the same. We move onto another topic. We can't compete with them."

0
0
0.000
avatar

Good morning @lnakuma

I just realized that I never actually thanked you for this amazing reply. Appreciate it a lot.

ps.
Would you mind if I take few minutes of your time? (I hope I'm not asking this question to often ;)

Anyway .... together with few core members of project.hope team - we'te trying to promote our recent publication: an article explaining economy behind our non-profit community project build on STEEM blockchain.

Perhaps I could ask you to spare few minutes and check it out and share your feedback with me.

I would absolutely appreciate it a lot. I read all comments and I drop solid upvote on each valuable one.

Link: https://steemit.com/hive-175254/@project.hope/3-ways-of-joining-our-efforts-project-hope-economy-explained

Yours, Piotr

0
0
0.000
avatar

I think that quantum computing is going to be highly disruptive as well. Some of the implications are truly mind-boggling.

0
0
0.000
avatar

Good morning @litguru

Would you mind if I take few minutes of your time? (I hope I'm not asking this question to often ;)

Anyway .... together with few core members of project.hope team - we'te trying to promote our recent publication: an article explaining economy behind our non-profit community project build on STEEM blockchain.

Perhaps I could ask you to spare few minutes and check it out and share your feedback with me.

I would absolutely appreciate it a lot. I read all comments and I drop solid upvote on each valuable one.

Link: https://steemit.com/hive-175254/@project.hope/3-ways-of-joining-our-efforts-project-hope-economy-explained

Yours, Piotr

0
0
0.000
avatar

Hey Piotr!

Nice recommendation of content again! Thanks!

Cheers!
Lucky

0
0
0.000
avatar

Dear Piotr,

Honestly I don't see a major growth in the amount of data or processing speed, still..

I do remember when few MB were more than enough, and now that I'm reaching 1TB mark and already deleted more than I wanted to, so I could have some free space for whatever is needed, already thinking about getting a new external drive of 1 or 2 TB says something..

( in a period of 20 Years, it's not that much, and processing speed I do believe is in direct co-relation to graphical "needs".. )

About A.I., there isn't really an A.I., just automation better coded than a few years ago, I was about to call it conjecture but it is not cause there are 'proof' of the non A.I. existence as known by the masses ( thank You Hollywood ), is it possible?! I would say, maybe!!

Now for what we really have, there is not a real need for tremendous processing speed, on the other hand a lot data is needed and fast access to it, tough there are quite a few things now-a-days that need a lot of processing speed it does not justify the explosive growth that the MSM talks about..

Anyway, thank You for the invite, it's always good to remember a few things and have this "talks" ;)

Take care my friend,
Cy

0
0
0.000
avatar

Just wait until optical computing replaces electrons with photons and data is stored holographically. That's more like the way the Star Trek computers operate.

0
0
0.000
avatar

Good morning @holovision

Would you mind if I take few minutes of your time? (I hope I'm not asking this question to often ;)

Anyway .... together with few core members of project.hope team - we'te trying to promote our recent publication: an article explaining economy behind our non-profit community project build on STEEM blockchain.

Perhaps I could ask you to spare few minutes and check it out and share your feedback with me.

I would absolutely appreciate it a lot. I read all comments and I drop solid upvote on each valuable one.

Link: https://steemit.com/hive-175254/@project.hope/3-ways-of-joining-our-efforts-project-hope-economy-explained

Yours, Piotr

0
0
0.000
avatar

Good afternoon, why do you need so much processing power? Where do you want to apply it, for the benefit of mankind or for the benefit of yourself? I understand that capacities are needed of course, but a small part of the population uses them.
Could you consider how you can use the huge computing power for the benefit of nature, for the benefit of those people who live in poverty?
It turns out to be a big gap between the possibilities of a group of people who have enormous resources, and those who do not even have bread at home.

0
0
0.000
avatar

Good morning @makedonsk

I believe I never actually thanked you for your reply.

ps.
Would you mind if I take few minutes of your time? (I hope I'm not asking this question to often ;)

Anyway .... together with few core members of project.hope team - we'te trying to promote our recent publication: an article explaining economy behind our non-profit community project build on STEEM blockchain.

Perhaps I could ask you to spare few minutes and check it out and share your feedback with me.

I would absolutely appreciate it a lot. I read all comments and I drop solid upvote on each valuable one.

Link: https://steemit.com/hive-175254/@project.hope/3-ways-of-joining-our-efforts-project-hope-economy-explained

Yours, Piotr

0
0
0.000
avatar

Hi! Piotr!

This is indeed a very interesting study and, I am very much elated that AI these days can now process data faster within 3 months. Now, I am not going to go through the whole TERMINATOR drama wherein humanity will be eliminated since robots and computers are now smarter than us.

As mentioned in the video, processing chips will be cheaper which in turn will make whatever computer/electronic based device cheaper. It would also ensure that everyone gets an access to better technology-e commerce, computer browsing- whatever content a person prefers to consume. Data aka information will be easily dessiminated and I think if it will be combined with the blockchain, it will create almost an airtight security. I am saying almost because somewhere, there might be a thing or two that needed fixing.

In regards to computer getting smarter than us, machines operate basically through logic. It does not know non tangible things like emotions or thoughts. In the end, they're just machines. Humans are creative and innovative. Machines cannot perform more than what they had been designed to do.

0
0
0.000
avatar

Hi @taskmaster4450!

Your post was upvoted by @steem-ua, new Steem dApp, using UserAuthority for algorithmic post curation!
Your UA account score is currently 6.138 which ranks you at #275 across all Steem accounts.
Your rank has not changed in the last three days.

In our last Algorithmic Curation Round, consisting of 98 contributions, your post is ranked at #7.

Evaluation of your UA score:
  • You've built up a nice network.
  • The readers appreciate your great work!
  • Try to work on user engagement: the more people that interact with you via the comments, the higher your UA score!

Feel free to join our @steem-ua Discord server

0
0
0.000
avatar

Cloud AI chips? Looks like a processing power will increase to service big IT corpos who are in business of surveillance and data harvesting. It doesn’t look like a processing power in favor of individuals.

0
0
0.000
avatar

Hello @taskmaster3350,

nice video and interesting topic!

Imo we're underestimating the speed at which things move in the whole AI area.

I guess as soon as we put more AI to work on specific AI problems we'll see a even more dramatic pick up in speed.

Ray Kurzweil is pretty much on point with his predictions for that last 30 years or so in technological advances and techno evolution.

But I believe we're in for some pretty dramatic surprises.

The horsepower as you called it on the AI processing capabilities front is an pretty interesting indicator of us slowly raising eyebrows.

Yes, us being slow is the key to understand why we're not ready in the slightest for this and we wont be for possibly a few generations even under normal circumstances.

But... the heck... there's Neuralink coming, right?

Hahaha!

Cheers!
Lucky

0
0
0.000
avatar

@crypto.piotr waiting for chinese mainland fabs. sha2 great to reach it.
ps. please no vids. no time to hear slow... text faster. regards. thx for the
0.003.

0
0
0.000
avatar

Summary:
Task discusses a report from Stanford University regarding the rapid advancements in artificial intelligence (AI) processing power compared to Moore's law. The report states that AI processing power is doubling every 3.4 months, which is significantly faster than the 18 to 24 months for traditional chips. Task emphasizes the potential implications of this rapid progress on various aspects of life, including automation, jobs, economy, and investing. He highlights the importance of keeping track of these technological advancements.

Detailed Article:
In this video, Task delves into a report from Stanford University that sheds light on the exponential growth of AI processing power in comparison to the well-known Moore's law. Moore's law stipulates that traditional chips double in power every 18 to 24 months. However, the report reveals that AI processing power is doubling at an astonishing rate of every 3.4 months, indicating a remarkable advancement in the AI industry.

Task emphasizes the significance of this rapid growth, particularly in the realm of applications. With processing power increasing exponentially, developers can create more sophisticated and powerful applications that require intense computation resources. This progression has led to significant advancements in computing over the past decades, fueling the digitization of various sectors.

The discussion touches on the potential implications of this rapid AI advancement, such as the acceleration of automation and its associated challenges like job displacement and income inequality. Task acknowledges these concerns but stresses the importance of embracing and adapting to technological progress rather than shying away from it.

Furthermore, Task addresses the debatable topic of whether Moore's law is slowing or coming to a halt, hinting at a potential new paradigm on the horizon. Companies like Intel, AMD, and Micron are already exploring alternative technologies such as carbon nanotubes to sustain the pace of technological advancement.

Task predicts that the proliferation of AI systems will lead to a pseudo-system where different AI technologies intersect, ultimately uplifting society as a whole. While acknowledging the current limitations of AI systems in terms of general intelligence, Task envisions a future where AI contributes significantly to various facets of life.

In conclusion, Task underscores the profound impact of rapid AI advancements on diverse areas such as investing, technology usage, financial markets, the economy, and employment. He encourages vigilance and engagement with these technological shifts, emphasizing their far-reaching consequences on society.

0
0
0.000