Please log in or register. Registered visitors get fewer ads.
Forum index | Previous Thread | Next thread
"Harness the Unsung Innovation Powering AI and Cloud Adoption" 10:09 - Feb 20 with 1071 viewsKeno

well do you find that Town has taken over you day to day thoughts?

That was the Title link of an email Ive just received and I immediately though it was about Marcus

Unfortunately its about how companies can use AI to improve business activity ....

"The acceleration of generative AI has spurred new industries and even reshaped our daily habits – triggering an exponential rise in data storage needs. As many facets of the economy become more AI-driven, we believe further growth hinges on the strength and scalability of data centres, which facilitate the storage, processing and analysis required to manage this surge.

We recently explored the factors keeping data centre demand high and why data centres and their broader ecosystem are likely poised for a multi-year secular growth run"


Poll: Should Hoppy renew his season ticket
Blog: [Blog] My World Cup Reflections

2
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 10:13 - Feb 20 with 1031 viewsRobTheMonk

I like nothing better than to sit back and have a can of Hyam-Bru.
[Post edited 20 Feb 11:25]
1
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:18 - Feb 20 with 920 viewsNthQldITFC

Data centres and data transmission are responsible for 1% of greenhouse gas emissions (as of 2023) apparently.

I've been out of the game for a while, but I'd assume that 'data centres' usage is primarily about storage and maintenance of data, whereas AI usage will be primarily about running processors.

In terms of the changes widespread AI usage brings I guess data will go from being relatively static (changed by human activity and simpler processes) to much more dynamic and in GPUs and CPUs rather than SSDs or old school HDDs? Although maybe load sharing will mean data is 'buffered' on SSDs etc for load sharing to keep customers happy? Overall stored data volumes probably won't change ever so much?

I would hazard a guess that the usage of electrical energy is going to be stepping up quite a bit with AI services being sold freely into all areas and scales of human businesses?

Anyone have an views on how much of a greenhouse gas impact this is going to have? Alongside any positive impacts.

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:40 - Feb 20 with 878 viewsSwansea_Blue

Very apt too. Harness the Unsung. He doesn't quite get the appreciation his performances this year merit.

Poll: Do you think Pert is key to all of this?

3
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:42 - Feb 20 with 880 viewsDanTheMan

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:18 - Feb 20 by NthQldITFC

Data centres and data transmission are responsible for 1% of greenhouse gas emissions (as of 2023) apparently.

I've been out of the game for a while, but I'd assume that 'data centres' usage is primarily about storage and maintenance of data, whereas AI usage will be primarily about running processors.

In terms of the changes widespread AI usage brings I guess data will go from being relatively static (changed by human activity and simpler processes) to much more dynamic and in GPUs and CPUs rather than SSDs or old school HDDs? Although maybe load sharing will mean data is 'buffered' on SSDs etc for load sharing to keep customers happy? Overall stored data volumes probably won't change ever so much?

I would hazard a guess that the usage of electrical energy is going to be stepping up quite a bit with AI services being sold freely into all areas and scales of human businesses?

Anyone have an views on how much of a greenhouse gas impact this is going to have? Alongside any positive impacts.


Data centres will be responsible for both processing and storage.

The biggest change AI makes in terms of power usage is that it requires GPUs rather than CPUs to run, and those can be much more power-intensive and require a lot of cooling. The initial training might be the worst of it but then running the models for people still takes GPUs for the most part. Storage does not come into it all that much, old school HDDs are not used outside of very long-term storage (for the most part).

Crypto-currency mining uses way more power than AI. The mining rigs are running GPUs all day, every day and at the end of it, you end up with something that has vanishingly small utility compared to AI.

https://spectrum.ieee.org/ai-energy-consumption

Going by what is being said there, Bitcoin just by itself already uses more than that (127 TWh per year compared to 29.3 predicated for AI).

There are lots of positives and negatives for AI. It's a tool that can be used for good or bad. It is still very early days for the technology (although the fundamentals have been around for decades).

With all that said, companies can and should be looking for ways to reduce the emissions of these datacentres. We do seem some novel approaches being tried. Cooling is usually the biggest issue with the datacentres as opposed to the electricity to actually the run the components.

I'm reminded a little of Project Natick that Microsoft has (had?) been running just off the coast of Orkney where they used the sea to cool the data centres.

https://en.wikipedia.org/wiki/Project_Natick

Poll: FM Parallel Game Week 1 (Fulham) - Available Team

1
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:55 - Feb 20 with 862 viewsKeno

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:42 - Feb 20 by DanTheMan

Data centres will be responsible for both processing and storage.

The biggest change AI makes in terms of power usage is that it requires GPUs rather than CPUs to run, and those can be much more power-intensive and require a lot of cooling. The initial training might be the worst of it but then running the models for people still takes GPUs for the most part. Storage does not come into it all that much, old school HDDs are not used outside of very long-term storage (for the most part).

Crypto-currency mining uses way more power than AI. The mining rigs are running GPUs all day, every day and at the end of it, you end up with something that has vanishingly small utility compared to AI.

https://spectrum.ieee.org/ai-energy-consumption

Going by what is being said there, Bitcoin just by itself already uses more than that (127 TWh per year compared to 29.3 predicated for AI).

There are lots of positives and negatives for AI. It's a tool that can be used for good or bad. It is still very early days for the technology (although the fundamentals have been around for decades).

With all that said, companies can and should be looking for ways to reduce the emissions of these datacentres. We do seem some novel approaches being tried. Cooling is usually the biggest issue with the datacentres as opposed to the electricity to actually the run the components.

I'm reminded a little of Project Natick that Microsoft has (had?) been running just off the coast of Orkney where they used the sea to cool the data centres.

https://en.wikipedia.org/wiki/Project_Natick


one of the joys of TWTD is how a random slightly stupid post can develop into quite a serious conversation and sometimes vice versa!!

Poll: Should Hoppy renew his season ticket
Blog: [Blog] My World Cup Reflections

3
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:31 - Feb 20 with 805 viewsNthQldITFC

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:55 - Feb 20 by Keno

one of the joys of TWTD is how a random slightly stupid post can develop into quite a serious conversation and sometimes vice versa!!


Apologies for the hijacking - I'm all pirate when it comes to the things I like to rant about!

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:35 - Feb 20 with 796 viewsNthQldITFC

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:42 - Feb 20 by DanTheMan

Data centres will be responsible for both processing and storage.

The biggest change AI makes in terms of power usage is that it requires GPUs rather than CPUs to run, and those can be much more power-intensive and require a lot of cooling. The initial training might be the worst of it but then running the models for people still takes GPUs for the most part. Storage does not come into it all that much, old school HDDs are not used outside of very long-term storage (for the most part).

Crypto-currency mining uses way more power than AI. The mining rigs are running GPUs all day, every day and at the end of it, you end up with something that has vanishingly small utility compared to AI.

https://spectrum.ieee.org/ai-energy-consumption

Going by what is being said there, Bitcoin just by itself already uses more than that (127 TWh per year compared to 29.3 predicated for AI).

There are lots of positives and negatives for AI. It's a tool that can be used for good or bad. It is still very early days for the technology (although the fundamentals have been around for decades).

With all that said, companies can and should be looking for ways to reduce the emissions of these datacentres. We do seem some novel approaches being tried. Cooling is usually the biggest issue with the datacentres as opposed to the electricity to actually the run the components.

I'm reminded a little of Project Natick that Microsoft has (had?) been running just off the coast of Orkney where they used the sea to cool the data centres.

https://en.wikipedia.org/wiki/Project_Natick


The usual feeding frenzy FOMO trigger referenced in there, which will of course eventuate and produce a lot of waste!

“I think the biggest responsibility is with institutions that are currently forcing AI on all kinds of solutions, regardless of whether it is the best fit, [because they’re] influenced by hype and fear of missing out,” he says. “It will be crucial to realize that AI is not a miracle cure and has its own limitations.”

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:48 - Feb 20 with 736 viewsKeno

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:31 - Feb 20 by NthQldITFC

Apologies for the hijacking - I'm all pirate when it comes to the things I like to rant about!


No need to apologise!!

Some of the best threads are ones that get hijacked and go on tangents

Poll: Should Hoppy renew his season ticket
Blog: [Blog] My World Cup Reflections

0
Login to get fewer ads

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:54 - Feb 20 with 683 viewsmo_itfc

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:18 - Feb 20 by NthQldITFC

Data centres and data transmission are responsible for 1% of greenhouse gas emissions (as of 2023) apparently.

I've been out of the game for a while, but I'd assume that 'data centres' usage is primarily about storage and maintenance of data, whereas AI usage will be primarily about running processors.

In terms of the changes widespread AI usage brings I guess data will go from being relatively static (changed by human activity and simpler processes) to much more dynamic and in GPUs and CPUs rather than SSDs or old school HDDs? Although maybe load sharing will mean data is 'buffered' on SSDs etc for load sharing to keep customers happy? Overall stored data volumes probably won't change ever so much?

I would hazard a guess that the usage of electrical energy is going to be stepping up quite a bit with AI services being sold freely into all areas and scales of human businesses?

Anyone have an views on how much of a greenhouse gas impact this is going to have? Alongside any positive impacts.


The point you've raised about data centers and transmission contributing to 1% of greenhouse gas emissions is an important one, highlighting the environmental impact of our growing digital footprint. As we navigate the complexities of AI integration into various sectors, it's crucial to consider both the direct and indirect implications on energy consumption and greenhouse gas emissions.

You're correct in differentiating the roles of data centers for storage and maintenance from the computational demands of AI, which primarily rely on processing power from GPUs and CPUs. The shift towards more dynamic data processing with AI does suggest an increase in energy usage, especially as AI applications become more widespread and computationally intensive.

However, it's also worth noting that advancements in technology often come with improvements in energy efficiency. For instance, newer data centers are being designed with sustainability in mind, utilizing renewable energy sources and employing advanced cooling techniques to reduce energy consumption. Additionally, AI itself can be a tool for optimizing energy use, not just in data centers but across various industries, thereby potentially offsetting some of its own carbon footprint.

The concern about AI services leading to a significant increase in electrical energy usage is valid. The integration of AI into all areas of human businesses does mean that the demand for computational power will rise. However, alongside this challenge is an opportunity for innovation in energy-efficient computing and the adoption of green energy solutions.

Regarding the greenhouse gas impact, it will largely depend on how quickly and effectively we can transition to renewable energy sources and improve the energy efficiency of computing technologies. As for the positive impacts, AI has the potential to significantly contribute to environmental sustainability by optimizing energy usage, reducing waste, and helping in the monitoring and management of natural resources.

In summary, while the expansion of AI and digital services will inevitably increase energy demand, it also offers unprecedented opportunities for making our energy systems more efficient and less dependent on fossil fuels. Balancing these factors is key to ensuring that the adoption of AI and the growth of data centers contribute positively to our environmental goals.

Written by ChatGPT.

Poll: Now that The Athletic has an article, is this takeover actually happening?

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:56 - Feb 20 with 655 viewsDanTheMan

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:54 - Feb 20 by mo_itfc

The point you've raised about data centers and transmission contributing to 1% of greenhouse gas emissions is an important one, highlighting the environmental impact of our growing digital footprint. As we navigate the complexities of AI integration into various sectors, it's crucial to consider both the direct and indirect implications on energy consumption and greenhouse gas emissions.

You're correct in differentiating the roles of data centers for storage and maintenance from the computational demands of AI, which primarily rely on processing power from GPUs and CPUs. The shift towards more dynamic data processing with AI does suggest an increase in energy usage, especially as AI applications become more widespread and computationally intensive.

However, it's also worth noting that advancements in technology often come with improvements in energy efficiency. For instance, newer data centers are being designed with sustainability in mind, utilizing renewable energy sources and employing advanced cooling techniques to reduce energy consumption. Additionally, AI itself can be a tool for optimizing energy use, not just in data centers but across various industries, thereby potentially offsetting some of its own carbon footprint.

The concern about AI services leading to a significant increase in electrical energy usage is valid. The integration of AI into all areas of human businesses does mean that the demand for computational power will rise. However, alongside this challenge is an opportunity for innovation in energy-efficient computing and the adoption of green energy solutions.

Regarding the greenhouse gas impact, it will largely depend on how quickly and effectively we can transition to renewable energy sources and improve the energy efficiency of computing technologies. As for the positive impacts, AI has the potential to significantly contribute to environmental sustainability by optimizing energy usage, reducing waste, and helping in the monitoring and management of natural resources.

In summary, while the expansion of AI and digital services will inevitably increase energy demand, it also offers unprecedented opportunities for making our energy systems more efficient and less dependent on fossil fuels. Balancing these factors is key to ensuring that the adoption of AI and the growth of data centers contribute positively to our environmental goals.

Written by ChatGPT.


I got to the second paragraph thinking "This reads very much like how ChatGPT writes".

Happy to know I can spot it!

Poll: FM Parallel Game Week 1 (Fulham) - Available Team

1
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 13:11 - Feb 20 with 631 viewsNthQldITFC

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:48 - Feb 20 by Keno

No need to apologise!!

Some of the best threads are ones that get hijacked and go on tangents


Don't get me started on tangents - what's wrong with a good old-fashioned British cosine?

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 13:15 - Feb 20 with 620 viewsNthQldITFC

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 12:54 - Feb 20 by mo_itfc

The point you've raised about data centers and transmission contributing to 1% of greenhouse gas emissions is an important one, highlighting the environmental impact of our growing digital footprint. As we navigate the complexities of AI integration into various sectors, it's crucial to consider both the direct and indirect implications on energy consumption and greenhouse gas emissions.

You're correct in differentiating the roles of data centers for storage and maintenance from the computational demands of AI, which primarily rely on processing power from GPUs and CPUs. The shift towards more dynamic data processing with AI does suggest an increase in energy usage, especially as AI applications become more widespread and computationally intensive.

However, it's also worth noting that advancements in technology often come with improvements in energy efficiency. For instance, newer data centers are being designed with sustainability in mind, utilizing renewable energy sources and employing advanced cooling techniques to reduce energy consumption. Additionally, AI itself can be a tool for optimizing energy use, not just in data centers but across various industries, thereby potentially offsetting some of its own carbon footprint.

The concern about AI services leading to a significant increase in electrical energy usage is valid. The integration of AI into all areas of human businesses does mean that the demand for computational power will rise. However, alongside this challenge is an opportunity for innovation in energy-efficient computing and the adoption of green energy solutions.

Regarding the greenhouse gas impact, it will largely depend on how quickly and effectively we can transition to renewable energy sources and improve the energy efficiency of computing technologies. As for the positive impacts, AI has the potential to significantly contribute to environmental sustainability by optimizing energy usage, reducing waste, and helping in the monitoring and management of natural resources.

In summary, while the expansion of AI and digital services will inevitably increase energy demand, it also offers unprecedented opportunities for making our energy systems more efficient and less dependent on fossil fuels. Balancing these factors is key to ensuring that the adoption of AI and the growth of data centers contribute positively to our environmental goals.

Written by ChatGPT.


"Them's some nice woolly word and generalisations, ChatGPT, how about some numbers and error bars?"

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

1
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 13:18 - Feb 20 with 603 viewsNthQldITFC

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:40 - Feb 20 by Swansea_Blue

Very apt too. Harness the Unsung. He doesn't quite get the appreciation his performances this year merit.


Powering Ali deserves his namecheck too, but I must admit I'm not familiar with Claude O'Doption - is he one of our Irish youth players?

# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
Poll: It's driving me nuts

0
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 13:24 - Feb 20 with 589 viewsblueasfook

"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 11:42 - Feb 20 by DanTheMan

Data centres will be responsible for both processing and storage.

The biggest change AI makes in terms of power usage is that it requires GPUs rather than CPUs to run, and those can be much more power-intensive and require a lot of cooling. The initial training might be the worst of it but then running the models for people still takes GPUs for the most part. Storage does not come into it all that much, old school HDDs are not used outside of very long-term storage (for the most part).

Crypto-currency mining uses way more power than AI. The mining rigs are running GPUs all day, every day and at the end of it, you end up with something that has vanishingly small utility compared to AI.

https://spectrum.ieee.org/ai-energy-consumption

Going by what is being said there, Bitcoin just by itself already uses more than that (127 TWh per year compared to 29.3 predicated for AI).

There are lots of positives and negatives for AI. It's a tool that can be used for good or bad. It is still very early days for the technology (although the fundamentals have been around for decades).

With all that said, companies can and should be looking for ways to reduce the emissions of these datacentres. We do seem some novel approaches being tried. Cooling is usually the biggest issue with the datacentres as opposed to the electricity to actually the run the components.

I'm reminded a little of Project Natick that Microsoft has (had?) been running just off the coast of Orkney where they used the sea to cool the data centres.

https://en.wikipedia.org/wiki/Project_Natick


My company are a Microsoft partner and we have migrated a lot of our stuff to Azure. We had a presentation last year on their data centres. Huge operations - massive rooms with rows and rows of racks. The undersea thing was funny yeah. Only downside is if it goes down totally. They basically just scrap the thing as it's difficult to replace a board in a rack when the thing is in a sealed unit under the sea. There is built-in redundancy to some extent but eventually the whole thing will fail and become unviable. So they will just basically bring it to the surface and replace it with a new one.

TWTD Veteran
Poll: How do you think season will end for us?

1
"Harness the Unsung Innovation Powering AI and Cloud Adoption" on 13:33 - Feb 20 with 559 viewsCoastalblue

The flipside of the conversion to a more digital world though is surely the fact that far less physical goods are manufactured and shipped now?

Media generally these days is streaming for most people, or at the least first downloaded in a digital format, compared to the records, CD's, cassettes, DVD's and even books that were once shipped around the globe before eventually being dumped in landfill somewhere that has to be a good thing?

No idea when I began here, was a very long time ago. Previously known as Spirit_of_81. Love cheese, hate the colour of it, this is why it requires some blue in it.
Poll: If someone promised you promotion next season, would you think

0
About Us Contact Us Terms & Conditions Privacy Cookies Advertising
© TWTD 1995-2024