Integration of the Doomstead with Dogchat is under construction.

Main Menu

The big AI revolution

Started by monsta666, Feb 20, 2024, 01:50 PM

Previous topic - Next topic

monsta666

Artificial Intelligence (AI) and large language models are all the rage these days and their impact is often touted as changing the world of work in a truly revolutionary way aka lots of job losses are likely to result from this technology.

These chat-based programs gain their knowledge by absorbing huge amounts of data and use that data to generate information that is often but not always correct. The results are quite interesting and if you have not already done so I would encourage you to start a conversation with one of these bots. The technology goes beyond texting however and can extend to generating pictures and even videos from simple prompts. In the next few years expect to see more and more AI generate pictures and even videos. You can possibly see which jobs are at risk just based on my descriptions and what nefarious uses this technology can be used for.

Anyhow, this thread will be dedicated to issues such as ChatGPT or other such programs and their potential issues they pose in the immediate future.

In the article I got linked up it goes on about the vast amount of energy required to power the various AI farms that have sprung to power these latest AI technologies. It sounds extreme, but due to sheer amount of electrical energy required there is a suggestion of powering future AI farms with nuclear energy.


Future data centres may have built-in nuclear reactors (BBC)

By Michael Dempsey
Technology reporter

A man with a shiny metal Concorde model on his desk, and old circuit boards on a shelf, is clearly in love with technology.

Chris Sharp is the chief technology officer at Digital Realty, a US business at the intersection of construction and high-tech. It builds data centres, the anonymous warehouses full of computers that keep the online world spinning.

And the advent of artificial intelligence (AI), which requires far more processing power than standard computing, has put rocket boosters under the data centre world.

His company has just built a huge new data centre in Portland, Oregon dedicated to AI. Just how different is this from an ordinary data centre?

"A normal data centre needs 32 megawatts of power flowing into the building. For an AI data centre it's 80 megawatts," says Mr Sharp.

AI systems are using all this extra electricity simply because they are doing so much more processing than standard computing. They are chewing through far more data.

Mr Sharp also points out that the entire web of technical support demanded by AI is greater. "You have five times more cabling, for instance."
A technician at a data centre in ChinaImage source, Getty Images
Image caption,
Data centres already used huge amounts of electricity, but AI is significantly increasing their power consumption

All of this points to a problem. How can AI grow when it requires so much more power to function?

Demanding ever more juice from the existing grid means competing with homes and other industries, and is not going to win the data centre sector any friends if blackouts result.

"Our industry has to find another source of power," Mr Sharp declares. He reckons that is nuclear.

More pressingly, he predicts that data centres in the not too distant future will come with their own dedicated, built-in nuclear reactors.

The technology in question is the much-touted Small Modular Reactor (SMR). These are designs for advanced reactors with about a third of the power generation of a traditional, large nuclear plant.

While there are currently no SMRs in commercial operation around the world, China is building the world's first, and similar technology is already used by nuclear-powered submarines.

Meanwhile, universities, such as the UK's Imperial College London, have for years operated small nuclear reactors for teaching and training purposes. Imperial's own reactor, located just outside London, was operational from 1965 to 2010.
The core module of China's forthcoming commercial SMR being transported last summerImage source, Getty

Today most companies developing SMRs for commercial use are focusing on helping towns and cities to keep their lights on. However, a clutch of specialist firms have decided that data centres are the best candidates for their SMR designs.

Dr Michael Bluck runs the Centre for Nuclear Engineering at Imperial College London. "Data centres are power hungry things, but with AI we're moving into a new level of power requirements," he says.

"There are about 50 SMR designs out there. The challenge is to build them in repeatable units, factory style, standardising production lines.

"There's no reason why a small fast reactor can't power a data centre, except that you have to get it past the regulator."

In the US, one SMR design from a company called NuScale has already been given the go-ahead by the Office of Nuclear Energy. Meanwhile, in the UK the Office for Nuclear Regulation is continuing to study SMR designs from Rolls-Royce and US tech firm Holtec International.

And US energy firm Westinghouse wants to build four SMRs in north east England, in Tees Valley, close to the existing Hartlepool nuclear power station.
Banner around links to stories about AI

Read additional stories on artificial intelligence
Banner around links to stories about AI

Dr Bluck thinks nuclear submarines offer a compelling safety case for SMR technology. "We build a reactor and put it in a submarine, and people sleep within two yards of it."

And he forecasts the tech sector's involvement in SMR development will prove decisive. "These guys have got oodles of cash!"

By contrast, Dr Doug Parr, chief scientist for Greenpeace UK, says that the high cost of SMRs will prove to be too great a barrier.

Unrealistic hype lies behind the cost estimates for SMRs," he says. "This hype will fall away as delays and difficulties emerge."

Dr Parr argues that data centre operators will get cold feet when revised costings emerge to render SMRs uncompetitive compared with renewable energy sources.

Greenpeace is also opposed to nuclear power on safety grounds, pointing out the risk of accidents, and the need to deal with the radioactive waste.

Spencer Lamb, chief commercial officer at British data centre developer Kao Data, doesn't view nuclear-powered facilities in the UK as being something imminent. "I've heard about SMRs, but it will take a long time to deploy a nuclear-configured data centre in the UK, and AI is happening now," he says.

Back in California, a company called Oklo says it has an SMR design that is almost ready to roll.

"AI is the catalyst, the main driver," says Brian Gitt, who is in charge of business development at the company.

Oklo plans to manufacture SMRs that it says can built both quickly and safely.

"People question the viability of nuclear power due to the waste product and risk of an accident," says Mr Gitt. "We recycle the fuel through our reactor multiple times, dealing with the waste, and the new reactors cannot melt-down, they are self-cooling and self-regulating."
An artist's impression of the exterior of an Oklo SMR and data centreImage source, Oklo
Image caption,
Oklo's designs for its planned SMRs appear to be going for more whimsy charm than your typical power station

Mr Gitt declines to reveal if any data centre firms have already signed contracts, but he says that "all the major operators are interested". He adds: "Power is their biggest problem, and they want to deploy this technology within four years, by 2028."

Mr Gitt says that his 25 years working in the energy business has convinced him that nuclear power is the only answer to the pressures the new AI data centre world faces.

"[Without us] they just don't have the power to turn on all the machines they need to have. We are signing letters of intent on specific data centre locations for the 2028 timeframe."

So what will a data centre's very own nuclear plant look like? Oklo envisages a large stainless steel tank that contains the SMR dropped into a 50ft (15m) deep hole in the ground. The data centre would then constructed on and around it.

Perhaps the ultimate tie up between SMRs, AI and data centres is evident on Oklo's board.

Sam Altman, the high profile leader of leading AI firm OpenAI has been its chairman since 2015. As Mr Gitt says of Mr Altman, "he identified this issue 10 years ago."

Thoughts anyone?

K-Dog

Your article turned into an article about nuclear power.  But it started off on topic.

The is a group of doomers I would love to introduce the Diner to.  I published this in another thread today.


I like the way they split the screen and the grid layout for the participants.  The right side of the layout was used to show info from chatGPT about the discussion as it progressed.  This must be chatGPT 4.0.  I use the free version 3.0.  Which is wrong as often as it is right.

I shudder to think of the power I have used with certain questions I have asked.

Will it put people out of jobs?

No, but capitalism will.

RE

Sadly, even the Black Widow probably can't stop AI from consuming still more Terawatts of power and infiltrating our existence, but she may at least slow it down a little.

https://www.politico.com/news/magazine/2024/05/22/scarlett-johansson-sam-altmans-washington-00159507

Sam Altman's Scarlett Johansson Blunder Just Made AI a Harder Sell in DC

RE

RE

They don't have enough Juice to keep the lights on reliably in Kenya, but the Big Gorilla will build a Geothermal plant just to power his Supercomputer AI farm there.

On the plus side, Come the Revolution, they could Nationalize the power plant, unplug the supercomputers and turn the lights back on.

https://africa.businessinsider.com/local/markets/microsoft-to-build-dollar1-bn-geothermal-data-centre-in-kenya-following-nigeria/52bsqm5

Microsoft to build $1 bn geothermal data centre in Kenya following Nigeria facility closure

RE

RE

Billionaires want AI, they just don't wanna pay for all the juice it takes to run the computers.  Somebody else should take the risk.

Who is out there demanding we have AI Data Centers?  Just tell the Billionaires, "You want this, you power it."  Has anybody seen any demonstrable improvement in their life from AI?  Most of the consequences so far seem to have been bad rather than good.  Deep Fakes and lost jobs for instance.

I see no pressing need for AI, and I certainly can't see why taxpayers would want to foot the bill.  Electric companies who might profit from it might want t, but they're already way behind on being able to supply power, so why do they want another headache?

https://www.datacenterknowledge.com/energy-power-supply/big-tech-wants-nuclear-power-but-doesn-t-see-role-as-investor

RE

K-Dog

#5
Quote from: RE on Jun 18, 2024, 12:51 PMBillionaires want AI, they just don't wanna pay for all the juice it takes to run the computers.  Somebody else should take the risk.

Who is out there demanding we have AI Data Centers?  Just tell the Billionaires, "You want this, you power it."  Has anybody seen any demonstrable improvement in their life from AI?  Most of the consequences so far seem to have been bad rather than good.  Deep Fakes and lost jobs for instance.

I see no pressing need for AI, and I certainly can't see why taxpayers would want to foot the bill.  Electric companies who might profit from it might want t, but they're already way behind on being able to supply power, so why do they want another headache?

https://www.datacenterknowledge.com/energy-power-supply/big-tech-wants-nuclear-power-but-doesn-t-see-role-as-investor

RE

QuoteI see no pressing need for AI,

All the talk now is hype.  A.I. was a dream and progress was painfully slow.  Anybody into computer science knows this.  Finally enough progress was made to get useful results.  But what kind of results really? 

A circular saw is a great tool if you have a wall plug to plug the cord into.  It is superior to a hand saw in many ways.  Even so it is not a perfect replacement, and at the end of the day the power saw is only a tool that gets a job done.  A wooden chair is a wooden chair is a wooden chair.  Power magic is not transferred to a product.  Power only makes production easier.  And A.I. has to be told what to do.

Besides not being magic, A.I. right now is like a precocious ten year old that has just consumed five candy bars and has been told to sit still.  Superior in many ways to humans, A.I. has serious shortcomings, and it will have a maturity issue for quite some time.

As for putting people out of work.  That is a function of the system, not technology.  Ask a true Luddite.  We don't hate technology.  We only hate technology stolen into private hands!

The private hands seem to be a problem.

RE

Quote from: K-Dog on Jun 19, 2024, 01:11 AMThe private hands seem to be a problem.

Private or Public, it consumes too much power.

RE

K-Dog

Quote from: RE on Jun 19, 2024, 09:07 AM
Quote from: K-Dog on Jun 19, 2024, 01:11 AMThe private hands seem to be a problem.

Private or Public, it consumes too much power.

RE

The public forced by common ownership to deal with reality could vote to limit profligate use to University Research.  All you get from private is chaos.

Could does not mean would, but it would be POSSIBLE.  With private it is impossible.

RE


RE

"We want data centers, but it can't be the Wild Wild West of data centers and crypto miners crashing our grid and turning the lights off."

"We" do?  Who is the we here?  Have they done a referendum asking people if they want a new Data Center in the neighborhood causing more frequent power blackouts and raising the rates on their electricity?  Everybody Knows we don't have enough power generation capacity to handle what the Billionaires have planned for us, but "we" want them so "we" have to pay for them.

And while big tech firms may be committed to using green energy, the Washington Post reports that, since they run on the same grid as typical power consumers, these data centers tend to eat up much of the renewable energy available, leaving power providers to source dirty energy to fill in gaps.

WHY do they run on the same grid?  Let them build their own fucking grid!  Put some solar panels on the fucking roof!  Power it with Green Hydrogen fuel cells!  Build a Fusion Reactor next door!

https://www.salon.com/2024/06/22/power-hungry-ai-boom-making-power-grids-dirtier-less-reliable/

RE

RE

You can see here why Capitalism makes reducing power consumption impossible.

Tech companies build a data center wherever there's a power company that says they can sell them juice.  Since the more juice they sell the more money they make, they all say, "Surre! Build it here!"

Once built, on those hot days they find they're short of juice, so they buy some from the grid.  Somebody somewhere fires up an old coal plant to make up the shortfall.

Regulations need to be put in place that require AI to have its own power supply independent of grid power.  Without that, the consumption will rise faster than renewables could power it.  True anyhow, but it would at least slow it down

https://cleantechnica.com/2024/06/24/weird-wacky-ways-of-powering-data-centers-are-under-consideration/

RE

RE

The tech sector's build out of data centers to support AI and the adoption of electric vehicles alone is expected to add 290 terawatt hours of electricity demand by the end of the decade, according to a report released by the consulting firm Rystad Energy this week.

Must...have...MOAR POWER!

Supplying the demand with reliable power is a matter of economic and national security, Womack said. Southern expects 80% of the demand through the end of the decade to be met by renewables, he said.

Which means that 20% more FF powered plants will be necessary as well, without retiring any of the currently running plants.  IOW, all of the added renewables does nothing to reduce carbon emissions at all, it just gets worse.

Opposite the banksters who see no solution to lack of affordable housing because of insufficient supply, the economistas tell us we must build terawatts of new power generation to feed the insatiable lust of billionaires for electricity.  Power for computers must be found...housing for the homeless cannot be.  The hypocrisy  of capitalists laid bare for all to see.  If they put half the investment they are making into data centers into affordable housing, the homeless crisis would be over inside of 5 years.

Still, not a single politician from any party including the minority parties has stood up to point this out and demand an immediate national affordable housing act.  Make a law that for every dollar Google spends on building data centers they have to match with a dollar spent on new housing.  All the housing and data center to be powered by the renewable energy plant they also have to finance.

Why is there nobody who will suggest this in any state assembly if not in CONgress?  Is there not 1 in the whole FSoA who can see the OBVIOUS?  Where is Captain Obvious when you need him?



https://www.cnbc.com/2024/06/30/failure-to-meet-surging-energy-demand-will-jeopardize-economic-growth-utility-execs-warn.html

RE

Knarf

I read an article this morning about how much energy it takes to run the AI programs. I will give examples in the article that convinced me to start using Duck, Duck, Go instead. I can't take part in the tremendous carbon footprint it is having right now. Maybe sometime in the future if they can bring the footprint WAY DOWN, but for now I can't use it with a clear conscience.

Today data centers run 24/7 and most derive their energy from fossil fuels, although there are increasing efforts to use renewable energy resources. Because of the energy the world's data centers consume, they account for 2.5 to 3.7 percent of global greenhouse gas emissions, exceeding even those of the aviation industry.
In 2021, global data center electricity use was about 0.9 to 1.3 percent of global electricity demand. One study estimated it could increase to 1.86 percent by 2030. As the capabilities and complexity of AI models rapidly increase over the next few years, their processing and energy consumption needs will too. One research company predicted that by 2028, there will be a four-fold improvement in computing performance, and a 50-fold increase in processing workloads due to increased use, more demanding queries, and more sophisticated models with many more parameters. It's estimated that the energy consumption of data centers on the European continent will grow 28 percent by 2030.
In 2019, University of Massachusetts Amherst researchers trained several large language models and found that training a single AI model can emit over 626,000 pounds of CO2, equivalent to the emissions of five cars over their lifetimes.

A more recent study reported that training GPT-3 with 175 billion parameters consumed 1287 MWh of electricity, and resulted in carbon emissions of 502 metric tons of carbon, equivalent to driving 112 gasoline powered cars for a year.
Once models are deployed, inference—the mode where the AI makes predictions about new data and responds to queries—may consume even more energy than training. Google estimated that of the energy used in AI for training and inference, 60 percent goes towards inference, and 40 percent for training. GPT-3's daily carbon footprint was been estimated to be equivalent to 50 pounds of CO2 or 8.4 tons of CO2 in a year.

Inference energy consumption is high because while training is usually done multiple times to keep models current and optimized, inference is used many many times to serve millions of users. Two months after its launch, ChatGPT had 100 million active users. Instead of employing existing web searches that rely on smaller AI models, many people are eager to use AI for everything, but a single request in ChatGPT can consume 100 times more energy than one Google search, according to one tech expert.


https://news.climate.columbia.edu/2023/06/09/ais-growing-carbon-footprint/

RE

Will be interesting to see how the court rules on this one.  Will Bezos get a free ride on the backs of the residential customers?  Place your bets here.

https://www.theregister.com/2024/07/03/utility_firms_launch_complaint/

RE

Knarf

Good catch. The energy wars will really get going now. Climate disasters plus, data centers are going to be fighting for energy. Scarcity in so many sectors now will also bring things to a boil. Fun times. :)