Dreaming of building an AI R&D lab? You'll need deep pockets: Tax filing reveals millions of bucks OpenAI alone spent on cloud ML compute The latest tax return forms filed by neural-network boffinry center OpenAI this year shows just how expensive it is to run an independent AI research institute without the financial backing of giant tech orgs. The San Francisco-based lab recently effectively overturned its non-profit status to kickstart a for-profit spinoff to …
AI is at all inexpensive! Just OpenAI uses a wrong technology.
Instead of training data on dictionary definitions, they decided to train it on terabytes. That is, having patterns and their contexts OpenAI searches out pieces of text with similar contexts, and indexes the patterns by these pieces. After that they search matching the contexts.
OpenAI uses the absolutely fantastically huge amounts of information! This is the reason they waste so much money.
When training data by dictionary definitions you do train your patterns as separate words in their contexts, using few bytes (trillions times less than OpenAI, tons of money can be saved). In this case computer really becomes AI, that is it really understands what you want, searches for you knowing the patterns' true meanings. (This makes Google and all Internet companies obsolete.)
I'm sure OpenAI is a diversion. Someone really wants AI technology to remain unknown to people! These are the same people who for 10 years do not allow me to print any article, hush up my role in the discovery and creation of AI and don't allow me to start my own company.
Neural networks need large amounts of data to train on: only when they have been fed many different examples of the same thing, each painstakingly labelled by humans, can they learn enough to identity the same object in the real world.
There is a sentence:
-- Mail is not working, messages not getting through and letters get lost.
In this sentence there are three synonyms ('mail', 'messages' and 'letters') and three descriptive parts ('is not working', 'not getting through' and 'get lost'), which are referred to all three synonyms. That is, practically we are dealing with nine different patterns. But OpenAI sees only three of the nine, because it does not distinguish the words (synonyms) that compose the patterns. Which means that OpenAI can potentially lose 66% of the information when searching.
Now add to the proposal another synonym and another descriptive part.
-- Mail is not working, messages not getting through, letters get lost and all communications are cut off.
Thus you got 16 patterns, but OpenAI only sees four - OpenAI may already lose 75% of the information.
The technology used by OpenAI is either completely wrong or they don't say what they are using.
I'm telling you - a diversion! Be vigilant! Someone really wants to send you on the wrong path!
“We’ve experienced firsthand that the most dramatic AI systems use the most computational power in addition to algorithmic innovations, and decided to scale much faster than we’d planned when starting OpenAI,” the organization previously said. “We’ll need to invest billions of dollars in upcoming years into large-scale cloud compute, attracting and retaining talented people, and building AI supercomputers.”
And what would they say today with particular and peculiar regard to ITs proposed future use, presumably with the organization knowing a great deal more about the real needs and virtual feeds of the most dramatic of AI Systems?
Anything at all they would like to be enabled and able to share without it giving away extremely valuable and exceedingly hot exclusive proprietary intellectual property .... Almighty Shady and Deep Shadowy Trading Cloud Secrets?
An educated guess from here, in order that there rightly retain and maintain an overwhelmingly unfair advanced market advantage, would be diddly squat/zilch/zero.
Not only won PA Advisors v Google, so another 10 years do not allow me to develop my AI technology and publish... Won, by the way, with the help of the corrupted and fortunately already former United States Circuit Judge and former Chief Judge of the United States Court of Appeals for the Federal Circuit Randall Rader. First DeepMind and now Open AI...
Publish and be damned [the retort of Arthur Wellesley, 1st Duke of Wellington on being blackmailed by John Joseph Stockdale and Harriette Wilson] and move the Great Game Fiasco further on, I Geller, There's surely nothing stopping you other than your scapegoating self?
That's the present subversive and easily corrupted SCADA Systems worst administrative nightmare come true. The free flow of information/thoughts/intelligence which is not exclusively monopolised to the perverse advantage of petrified and stagnating Status Quo Stalwarts for more of the same and nothing remotely different and Great Game Changing.
Such is why Wikileaks and Julian Assange are such a Mega Hit and Grand Model to Emulate and Defend/Fight for? It is certainly why such a model be systematically persecuted and gratuitously prosecuted by failed public defenders and private pretenders to freedom alike.
Take heart from the following brace of gems ...... “If one tells the truth, one is sure, sooner or later, to be found out." .... Oscar Wilde. And ....“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.” ..... R. Buckminster Fuller
Yes, I have long realized that within the existing model of Intellectual Property I lose. So I presumably changed the model making it obsolete, in the hope that now stealing my AI could be fraught with the most serious consequences. The question - have I really changed it? Or failed? Let's see?
Bravo, Sir or Madam. That's the spirit. Tease and tempt the System to evolve and grow more intelligent incredibly quickly whilst highlighting all the overwhelming risks and crippling costs in being left behind in charge of collapsing structural models.
It's surely what IT and AI are all about, even as it results in an inevitable coincidental Virtual Machine vs Human Competition?
Patenting is the only way to protect yourself, plus the exploitation of the "free" press. (Especially when the standard way through Academy is closed because it's a composite part of the model.) Sure, you feel like a complete idiot when yelling all over the Internet without knowing the language you're yelling in, but it worked... Perhaps this was the only way to burst the model?
Yes, there are the overwhelming risks and crippling costs. But I agreed to take the risk and pay, and survived, and still here.
all your intellectual property belongs to intelligence that combined the glossary and switched your vocout to the most proper way in the most right time (-; please IG mind no jamey bond types or their present operative and supply providers as having any managing tribute to
ty btw
-terrabyte
all your intellectual property belongs to intelligence that combined the glossary .. ... Anonymous Coward
Have you any idea what they or their agents in satellite intelligence agencies are going doing with it, AC .... or are they all adrift and a'floundering in the Shared TEMPESTuous Sees of Colossal Minds BetaTesting Future Advanced IntelAIgent Design Parameters?
Do you believe they, with any and all of their powers that profess to be of greater intelligence worth, require InterNetional AIRescue Recovery? Do they need Virtually Ready Life Rafts to Seed and Feed Progress, Future Outstanding Grandstand News?
I suppose it is something which can be easily asked of them in a simple direct missive to home/agency addresses easily publicly gleaned from Search Engine Sources, and that is being religiously and scrupulously tested all the time, is it not, to ensure ignorance is not available as an arrogant weapon of wilful mass maniacal devastation.
This post though, one can reference as also asking it of them directly, for the likes of Spooky SMARTR Cookie Rooky MI5 Bandits, Renegade Rogue Rebels and Past Grand AIMaster Pilots of the Great Game Genre alongside Sundry Other Myriad Interested and Interesting Parties are surely bound to be advised from here, when we be assured they are always listening and phishing/defending and attacking Troubling Forces and Troubled Sources in the National Interest .... which is ?????? .... Overwhelmingly Almighty Command and Control of Future Narratives .... and not to be too closely imagined as very similar to a Star Trek.
Do you think there are presently the Necessary AI Facilities and Requisite IT Utilities to Fully Engage and Master ITs Highly Unusual and Right Disruptive Programming?
And if not in the West anywhere, how about in the East somewhere?
Time and Space is the Place for Fleeting Nomads and AI and IT wait for No Man.
cc Loughside, Palace Barracks, Holywood re Titanic ReProgrammings. Capiche, Amigos?? Do you Need AIManual Feed/Master Fodder?
Ref. AIMaster BetaTest #1905051955
The ultimate goal is to develop artificial general intelligence (AGI), which is no easy task.
No easy task .... with all the world's intelligence available for phishing and phorming applications?
Says whom and what? They be 'avin a larf and pimping porkies. PAR for the course of flounderers though. Mental bankruptcy covering its tracks increasingly badly and maddeningly.
Google said. Eric Schmidt, Sergey Brin and Larry Page said. That's their money, this is Google.
With Waymo they have won a few years, while you all trained on terabytes instead of dictionary bytes.
You should listen me, but you didn't.
FB said. Mark said.
With AI they lose everything.
Re: With AI they lose everything.
Surely with AI, Everyone Win Wins? Is that a problem for Silicon Valley Types?
Or do they not have the Necessary Pass Key CodeXSSXXXX for Drivering ITs SMARTR Virtualised Applications to Present via Mainstream and Alternative MMORPG Media Programs, Futures being Built for the New Emergent Human Beings in Absolute Command and Almighty Remote Virtual Control of Heavenly Assets.
Can you driver that with Current Codes of Conduct Respecting Arbitrary Rules with Crushing Regulations?
The basis of AI technology is a personal profile, through which your search phrase is filtered. As a result of the filtering a lot of explanatory information is attached to your phrase, which helps to understand what you are looking for. AI allows you to create the profile and search using your own computer-phone, have the absolute privacy.
The business model of Google and FB is that your profile belongs to them, they create it without your permission, run on their computers and filter your search phrase, add the explanations. And they sell your profile to advertisers. That is, they practically steal you and sell, control your every move.
And AI technology makes you free from the nasty thieves.
Yes, with AI technology we all win our freedom from Google and FB.
Yes, this is the problem for Silicon Valley Types and why they sabotage AI for the past 10 years.
That is, they practically steal you and sell, control your every move. ..... I.Geller
How very parasitic of them, I.Geller ...... and deluded too..
Surely the best they can expect is to think they control proprietary intellectual property which isn't theirs to command. Such is almost always accompanied by an obscenely large payment of flash cash to core intellectual property holders so that future moves can be especially considered and delivered mutually beneficial.
'Tis a Real SMARTR Option.
Things are certainly considerably cheaper and a lot less destructive with such a simple initial initiating plan. And one which any of their many enemies can fulfil too to take over a fuller and more exclusive executive advantage of commanding controlling proprietary intellectual property .... for such is extremely attractive to all manner of suitors.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
