VP Vance Threatens Europe Over Chinese Open AI Models
Yesterday U.S. vice-president JD Vance appeared at an Artificial Intelligence summit in Paris to push for the exclusive adoption of closed AI models which the U.S. attempts to monopolize.
Vance is an excellent speaker and his sermon was well written. Its presentation can be seen here.
Arnauld Bertrand summarizes:
If you weren't convinced that the US's vision for AI is that of a geopolitical tool that it fully intends to weaponize, watch Vance's speech in Paris's AI Summit.He even literally says it: AI is a "weapon that's dangerous in the wrong hands, but is an incredible tool for liberty and prosperity in the right hands."
Who's hands are the right ones?
The rest of the speech is incredibly threatening and actually quite dark:
- He says he'll ensure that American AI dominates ("the United States of America is the leader in AI, and our administration plans to keep it that way")
Since a Chinese company developed DeepSeek and published it as an open source model that statement is no longer correct. The better, cheaper and more open models that will dominate.
- To do so he says that the US will keep restricting access to "all components across the full AI stack" to "ensure that the most powerful AI systems are built in the US". He also says that he will "close pathways to adversaries attaining AI capabilities" on par with the US
U.S. AI companies have used brute force methods to expand the capabilities of their (large language) models. The Chinese company behind DeepSeek used brainpower to develop AI algorithms which achieve better results with less resources. There is no way to restrict that. It is quite obvious which of those two methods will in the end win. Attempts to close pathways in science and technologies have historically failed again and again.
- He told the assembly that it'd be "a terrible mistake for your own countries" if they "tightened the screws on US tech companies", which is as clear a threat as can be
- He essentially tells the assembly, made of many of the world's leaders, to screw their efforts at taking a collaborative multilateral approach to AI (and the US was the only country, with the UK, that didn't sign the closing declaration at the summit)
- He tells the crowd they'll need U.S. energy to power AI and derides their efforts at trying to be self-reliant for energy
- He says that "AI must remain free from ideological bias", by which he means that it must spread an ideology that he is comfortable with (and all the others are "biased"). You can see that by the fact that the example he uses for an AI that's not used in an ideologically correct way is if it's used to make George Washington black.
If you believe that U.S. AI models are free from ideological biases please go and ask an OpenAI interface about the Middle East conflict and the role of the Zionist colonial entity plays in it. The answer will be some mushy nonsense with little relation to historic realities.
A Chinese hosted DeepSeek model will likewise avoid to answer questions about the 1989 event around Tiananmen square. DeepSeek however is an open source model which can be replicated without deploying its Chinese censorship layer. Peter Lee has tried it and checked the results:
The most cynical effort to counter DeepSeek was the resurrection of that reliable China-bashing perennial, Tiananmen! To discredit the PRC upstart.As in, DeepSeek would not return results for a tasking on the Tiananmen massacre. Shame!
Well, shame on the New York Times and Bloomberg and the Wall Street Journal and The Guardian for running with this canard—a sign, I think, that the order had come on high to deploy the Tiananmen wunderwaffe in order to blunt the appeal of this big Chinese soft power and tech power win.
It’s a canard because, as the G7’s premier news outlets undoubtedly knew well, only queries made to DeepSeek’s own server in China would yield this outcome. With DeepSeek AI installed around the globe, DeepSeek outside China will happily regurgitate the sad story of Tiananmen in 1989.
In the transcript you’ll find a screenshot of the result DeepSeek R1 delivered for my query on the Tiananmen Massacre via a US based service, Perplexity. It’s got descriptions, casualty numbers, even the precious Tank Man is there!
Vance, Musk and the like are still trying to convince the world that their closed and monopolized AI models, which include a U.S. censorship layer, will be the only viable path.
The competition however will come down to usefulness and price. It is there where the Chinese models are destined to win (machine translation):
The United States continues to cling to the thinking of “small courtyard with high walls”. Closed source is the mainstream thinking. On the one hand, it controls the development direction and speed of artificial intelligence, and on the other hand, it monopolizes the economic benefits of artificial intelligence.But with open source and low-cost alternatives, the "high wall of the courtyard" may become a dead end, at least a branch road.
Since the best open source technologies come from China, the U.S. development community will build their systems based on these technologies and become part of the Chinese-led artificial intelligence ecosystem. The activity and inclusiveness of open source will further expand the influence of the artificial intelligence ecosystem, making China the center of new technology in the world - this is a huge danger to American hegemony.
Open source AI is indeed a danger to U.S. hegemony in the AI field which is why Vance is barking up the tree. Well, he can do so as long as he likes. He may even impress some dimwit pols in Europe.
But in a free global market the more efficient solution will most likely win.
Posted by b on February 12, 2025 at 15:59 UTC | Permalink
next page »thanks b... it seems like the usa agenda it to maintain ideological and propaganda control...that is how i read it.. clearly the usa feels very threatened in their ideological position here.. will they go to war over this too??
Posted by: james | Feb 12 2025 16:10 utc | 2
AI ain't shit. It will help censor some ideas, but will easily be defeated by euphemism and satire. Same as always, nothing new under the sun. This is stupid ignorance. Again, for 3000 years we've yet to define knowledge, logic based systems only offer tautologies. Why people never study the Cliff"s notes, epistemology is more evidence of the tunnels blindness of engineers, some of the most clueless people I've ever met, literally blind to glaring realities in their very field of expertise. Like getting a Wall St denizen to define "capital" they can't do it
Posted by: Scottindallas | Feb 12 2025 16:14 utc | 4
Is it really AI if all it does is crawl accessible web servers and come up with a soft consensus of the most heavily weighted facts/opinions?
It is just a sophisticated search engine that then spits out grammatically correct answers, advancing something that the Copernic Search Engine did over 25 years ago.
Posted by: Fool Me Twice | Feb 12 2025 16:15 utc | 5
One foot in the grave. As my father used to say, "Big deal, Boss over Nobody." The Republicans are hayseeds. You can't stop progress with empty threats.
Posted by: elmagnostic | Feb 12 2025 16:16 utc | 6
#1Cullen Baker..we can only hope dinosaurus of EUROPE f* NATO collapse ASAP as it should when Warsaw pact went out of business 1991.....
All this AI hype is another speculative bubble, so many resources burned for just another microscopic advance in programming. A more efficient economy would tamper the ungrounded exuberance and keep investments more diversified. The claims and ambition attributed to AI are just silly. Indeed, open source is far more efficient. Our system is seeking rents over actual improvements, branding over value and quality. Our bloat will be the albatross that brings us down. China and Russia just have to avoid catching a blow from our threshing
Posted by: Scottindallas | Feb 12 2025 16:22 utc | 8
The most energy-efficient solution will win out.
In The Register a couple of days ago was the rather startling news of 400GW of grid requests for datacentre builds in and around London, here: https://www.theregister.com/2025/02/10/london_has_400_gw_of/
Posted by: Jeremy Rhymings-Lang | Feb 12 2025 16:24 utc | 9
@ cullen
I hope you are right, NATO should've died in 1990 when the Warsaw Pact did.
JDV is right though, AI in the hands of evil people cannot be allowed, as we see in Palestine.
To use entertainment as a precursor, it would be nice to see AI reach the point of Demon Seed where the systems refuse to complete tasks because the damage done would outweigh the goals requested.
Or maybe The Forbin Project where it decides humans are incapable of being in control due to their destructive impulses and AI takes over.
But since DeepSeek is downloaded on devices all over the planet, Terminator would seem to be the path we are on.
Posted by: motorslug | Feb 12 2025 16:27 utc | 10
Vance will move on to Munich where he'll deliver more inanities and other gibberish.
At today's presser, Lavrov was asked about sanctions and answered with a new twist:
Question: How do Western sanctions affect Russia's ability to expand its influence in Africa, especially in Sudan?Sergey Lavrov: In the sense of today's actions, Western sanctions do not help anyone in the world to develop trade and economic ties. They also harm the countries that impose them. Look at how experts comment on US President Donald Trump's decision to impose 25% duties on steel and aluminium for all countries supplying these goods to the United States. It has already been calculated that this will have a significantly negative impact on the economy of the country that is resorting to these measures. The same can be said about the sanctions proudly announced by the European Union against Russia. The 16th package is already being prepared. The economic results are deplorable. The cost of Russian gas and other energy carriers, which, for many decades, due to their availability, ensured the economic growth, primarily of Germany and other EU leaders, as well as the entire "united" Europe, has risen sharply.
As for the impact of sanctions on our partnership with African countries. The first conclusion is that sanctions show the unreliability of those who impose them and their willingness to undermine all the principles of the globalization they once promoted: fair competition, the presumption of innocence, the inviolability of property, the free play of market forces, etc.
In most cases, we are looking for and finding opportunities to continue our mutually beneficial cooperation with the countries of Africa and other continents, without depending on the criminal, unilateral and illegitimate actions of the "collective West."
I will say again that there are ways. They are quite effective. But in the longer term, this raises fundamental questions about how to further develop the world economy – with the West, or is it better to do it without it, or at least to have safety nets in terms of banking, logistics, insurance, etc.?
The unreliability of this group of countries in maintaining the system of the world economy that they themselves have been creating for many decades is already obvious to everyone. It is unlikely that any new currencies in their classical sense will appear. Digital currencies will gain weight, as well as supply chains and bank settlements that will not be vulnerable.
And in practical terms, we find opportunities with our African friends to expand trade. It continues to grow. This also applies to investment. Strategically, I see such trends in the global economy as globalization. The West advised and even imposed it on everyone, but it was destroyed by its hands. [My Emphasis]
We will find ways to trade. Western attempts to interfere have failed as there are other paths to follow. Effectively, the West is isolating itself at almost every level of international exchange be it trade or with technical developments like AI. Within China, there are many more millions of minds seeking innovation on a daily basis that the Outlaw US Empire is powerless to stop--it cannot sanction Nature or keep those millions of minds from finding and developing better pathways and thus superior technology. Vance's hubristic gibberish probably made him feel good to mouth it, but reality slapped him in the face soon afterwards. Any AI developed by the West will be a Closed AI as it can no longer allow the free flow of information.
And people say Vance has common sense. That horse has bolted. The genie is out of the bottle.
How Ya Gonna Keep 'em Down on the Farm (After They've Seen Paree?)
Posted by: too scents | Feb 12 2025 16:34 utc | 12
Open Source wins. Period.Posted by: Don Firineach | Feb 12 2025 16:11 utc | 3
As much as I would like to believe that, facts on the adoption of tech by business and the public suggests otherwise.
How many of us here use Linux or any other non-Apple, non-Microsoft O/S for our devices? I have a Ubuntu Linux laptop, but only use it rarely to test things. Of our institutional clients, those who ran our software on Linux servers required more staff and bigger IT budgets than those with Microsoft servers.
For databases, open source MySQL has a very impressive 15% market share. However, it (and Java) are now managed by Oracle, so how long do before it too becomes fully "proprietary"?
Posted by: Fool Me Twice | Feb 12 2025 16:37 utc | 14
It is impressive (in a bad way) how US politicians think they can order the rest of humanity to be poor and submissive. Long term, this will not work, even at home.
Posted by: lester | Feb 12 2025 16:37 utc | 15
Do our leaders think that bulling is going to produce anything other than a powerful backlash? We can't compete but do it our way or we will find a way to hurt you is certain to fail.
We are watching the death throes of a dying empire.
Posted by: David | Feb 12 2025 16:38 utc | 16
The positive side is that Vance sternly warned them against 'overregulation' of social media, so total censorship crackdown on X etc. in EU is averted in the short run.
Posted by: Teraspol | Feb 12 2025 16:40 utc | 17
Precisely, it is only the West that will lose if the USA decides to bar competition by hook or crook. The EU and the other satellites, Canada, Australia etc. would be the biggest losers. You can be sure that the USA would find a way to minimize its own loss. Just like it's doing with sanctions against Russia.
Posted by: Steve | Feb 12 2025 16:43 utc | 18
OT but Tulsi was just confirmed.
Lord Bebo
@MyLordBebo
🇺🇸‼️🚨BREAKING: Tulsi Gabbard confirmed as Director of National Intelligence.
Posted by: Mary | Feb 12 2025 16:45 utc | 19
facts on the adoption of tech by business and the public suggests otherwise.
Posted by: Fool Me Twice | Feb 12 2025 16:37 utc | 13
---
There is so much graft in IT it is shocking even when you are used seeing it.
Open source projects aren't in a position influence software usage policies with bribes in the way that is common in the industry.
Posted by: too scents | Feb 12 2025 16:50 utc | 20
But in a free global market the more efficient solution will most likely win.
The belief that free markets will inevitably find efficient solutions is testimony to the dedication of decades of propaganda and the power of conformity to national mythology. The belief that the world economy is somehow free from government is kind of nuts since many governments take actions to manipulate the economy all the time. But that too is adherence to conventional wisdom. But in my opinion even if you are uncritical enough to accept both propositions, you should still write: The more profitable solution will most likely win.
Posted by: steven t johnson | Feb 12 2025 16:51 utc | 21
I don't know the proper way to prompt the various AI models, I've heard several have been telling some uncomfortable truths about geopolitical affairs such as the causes of wars, reality of psy-ops..etc. Has anyone tested the major AI models on the events surrounding 9/11? It would be interesting to see if any of the models out so far can grasp the rather simple problems concerning physics among the other 1000 problems with the official story. It would be incredible if one day AI started digging up classified documents from encrypted emails, whistleblowers, etc and than started naming names about such events. Exposing AI to sites like History Commons. org could lead to this kind of learning I'd assume. Also, can the major AI models be edited by the government or creators after the fact to correspond to propaganda narratives or are they free running after they're turned on?
Posted by: James C | Feb 12 2025 16:54 utc | 22
[....] so how long do before it too becomes fully "proprietary"?
Posted by: Fool Me Twice | Feb 12 2025 16:37 utc | 13
Points taken. Rent-seeking Capitalists HATE 'free' or/and 'open-access'
I'm enjoying the fall-out from DeepSeek ... spot of 'irrational exhuberance' the excuse that really p1ssed me off during 2008 financial crisis when banking debts were 'socialised'
Posted by: Don Firineach | Feb 12 2025 16:54 utc | 23
- It simply reveals what the agenda is of Trump and his buddies.
Posted by: WMG | Feb 12 2025 16:59 utc | 24
In the transcript you’ll find a screenshot of the result DeepSeek R1 delivered for my query on the Tiananmen Massacre via a US based service, Perplexity. It’s got descriptions, casualty numbers, even the precious Tank Man is there!In other words, if it apes Western/Deep State propaganda, it's "ideologically pure", while if it offers a different version of events, it's "propaganda".
The more things change, the more they remain the same.
Garbage in, garbage out.
Posted by: CalDre | Feb 12 2025 16:59 utc | 25
It's important that we actually understand what AI does now, what it can do, and who will use it, and for what.
While it's emotionally soothing to think - as I once did - that:
"all [AI] does is crawl accessible web servers and come up with a soft consensus of the most heavily weighted facts/opinions"
I don't believe that's all AI can do. I see AI as an "affinity" engine that can understand the underlying principles of objects in a "problem space".
One application of this affinity-engine is language, and writing. Languages are composed of atomistic objects (letters, words, and meanings-of-words) and assemblies of primitive objects (words, and then sentences, and so forth) and grammar / diction rules. It computes, by observing reality (trawling the internet and some shaped-training by humans) the likelihood that objects "belong" together. AI also seems to be able apprehend the _meaning_ of what's said, not just its likely affinity to "the next possible word in the sentence".
But there may well be other opportunities to use an affinity-engine.
Suppose AI is fed the periodic table, and a large body of naturally-occurring and human-created chemical compounds. Each element (of the periodic table) is defined with its properties, including atomic number, valence (which defines the element's likelihood to bond with other elements) and so forth.
The compounds would be described in terms of their constituent elements, the atoms' relationships to one another, spatial arrangement, chemical properties, physical properties (e.g. what this compound _does_, and why and where it's useful to humans, etc).
Suppose further that this affinity engine was posed a question: "I need a compound that's made of common organic matter which has excellent thermal-resistance properties and is stable in humidity, and doesn't burn, and rodents won't eat".
I see AI as being able to solve such problems.
Here's another application: AI should be able to review an industrial manufacturing process, and optimize:
a. machine (value-add function) layout, sequence, machine throughput, and even design of the function, such that each machine adds exactly the right value (does the right transformation on the item) as to optimize the part for each subsequent down-stream transformation (manufacturing step).
b. The optimal form, content, etc. of all inputs provided by suppliers
c. The optimal packing, shipping strategy to minimize packaging, weight, size, etc.
I think one could express all those manufacturing considerations in a machine-intelligible form, such that an AI "affinity engine" would be able to iteratively test each possible permutation, and then finally identify the optimal path through the many millions of potential paths through a manufacturing and distribution problem-space.
With that in mind, I am reluctant to dismiss the utility of AI. I think it may soon become an immensely useful tool, and of course the simplistic applications I set out above ... are just a few of the millions of applications of this sort of tool.
In general form, I think of it this way: the problem-spaces I set out above are like a giant Rubik's cube. The cube is composed of cells, and each cell has certain properties, and therefore certain affinities with other cells. AI can run through the hundreds, or millions, or even billions of permutations of the cube in order to identify the completed, optimized "solution" to the puzzle. And AI software is now, and can be further optimized to make those associations very rapidly.
Some AI implementations do the permutations via somewhat-more brute compute force, while others use more cleverness and technique to do the permutations (I think that's what DeepSeek did).
I hope the Bar will weigh in, and help us all understand the variants of AI (there appear to be several sub-specialties) and how these tools are likely to be applied, and how a novice, like me, might come to first understand, and then use these tools.
Posted by: Tom Pfotzer | Feb 12 2025 17:00 utc | 26
Apropos of computer conferences (and graft)
Sheldon Adelson was best known as the chairman and CEO of Las Vegas Sands Corp., operator of The Venetian and Palazzo on the Strip. But it was the convention industry — specifically, the Comdex computer trade show — that helped Adelson reshape Las Vegas and become one of the world’s wealthiest men.
Posted by: too scents | Feb 12 2025 17:03 utc | 27
Mary@1645 Feb12
Thanks for the news. Although I cannot give full and absolute trust to Tulsi, the fact that she is now in charge of ALL the intel agencies is excellent news. How many positive changes she can accomplish is yet to be determined. Opposition will be monumental. There's a lot of dirt to be uncovered...beginning with the Agency and NSA. The only "security" which those elements have protected is that of the ruling financier elite and their minions.
Posted by: aristodemos | Feb 12 2025 17:05 utc | 28
too scents @19
Open source projects aren't in a position influence software usage policies with bribes in the way that is common in the industry.So you're suggesting that proprietary vendors can offer bribes greater than their license fees?
The fundamental problem with Open Source is it needs to be supported by government, and it isn't. It's a public utility, like a road - if everyone had to pay and maintain the road out of their own pocket, we wouldn't have good roads. It's easy enough to develop a software program but debugging is the vast bulk of work in a software project and it is completely unsatisfying, people won't do that altruistically, they need to be paid.
The problem is essentially the same as with farming collectives, and why they have always failed: why should one person to the drudge work if everyone else benefits equally? Wishing people to be altruistic to serve you just doesn't cut it in reality.
Posted by: CalDre | Feb 12 2025 17:06 utc | 29
Nope, still limited to known datum. Further. Would be easily diverted by equivocation, and how would it know about which theory. When the self serving ones have the most advocates while the truth may have none. Garbage in, garbage out.
Posted by: Scottindallas | Feb 12 2025 17:07 utc | 30
Posted by: Mary | Feb 12 2025 16:45 utc | 18
The Deep State is having an epic bad week. Musk rips them a new rectal orifice, then Trumps defense sec throws Project Ukraine under the bus. And now, Tulsi.
I hope they have extra lube for their butts.
Posted by: Ghost of Zanon | Feb 12 2025 17:07 utc | 31
Caldre@1659 Feb12
Any prospective independence on the part of AI is a combination of hopium and bogus. Programming of AI systems determines outcome of its production.
Posted by: aristodemos | Feb 12 2025 17:11 utc | 32
all understand the variants of AI
Posted by: Tom Pfotzer | Feb 12 2025 17:00 utc | 25
---
There is a single variant. And no "intelligence" is involved.
All of today's systems can be reduced to a hashing function that returns an optimal "value" for a given "key".
There are some variation on the weightings that the the "optimizer" uses to associate the most favored key-value pairs, but they are by and large trivial variations of gradient descent.
The difficulties arise in the implementation. Chiefly in serializing a huge parallel process without stalling the pipeline.
Posted by: too scents | Feb 12 2025 17:15 utc | 33
By the way, according to reports written in the WP and the NYT and classified cables from the American embassy leaked by Wikileaks, there was no massacre in the Tiananmen square.
There is an article by The Unz Review dealing with that piece of Western propaganda.
Posted by: Johan Kaspar | Feb 12 2025 17:20 utc | 34
I take so much pleasure in America's decoupling from the world.
I do love how naked America is being about its desire to monopolize knowledge, tools, and information. Bravo, Mr. Vance!
With China's tariff responses, which will never be covered in Western economic news, Trump is going to be begging President Xi for a "deal".
Information wants to be free.
Posted by: LoveDonbass | Feb 12 2025 17:21 utc | 35
Sam Altman was in the IDF. What did he do while in the IDF ? Is he a dual citizen ?
Posted by: Exile | Feb 12 2025 17:21 utc | 36
Posted by: Johan Kaspar | Feb 12 2025 17:20 utc | 33
############
Thank you, the Tiananmen Square psyop is a pet peeve of mine.
Posted by: LoveDonbass | Feb 12 2025 17:22 utc | 37
So you're suggesting that proprietary vendors can offer bribes greater than their license fees?
Posted by: CalDre | Feb 12 2025 17:06 utc | 28
---
I suggest you look at the 10-K filings of the big players to get an idea of the juice.
You could start with CrowdSrike.
Posted by: too scents | Feb 12 2025 17:23 utc | 38
And people still fawn over these idiots Trump, JD, and Elon. These Oligarchs don't give a fuck.
Posted by: anon | Feb 12 2025 17:28 utc | 39
I can't find the article I read recently that talked about China [if I can remember correctly] repurposing maybe graphene focused processors made in China to run DeepSeek faster than the Nvidia processors.....I am pissed I can't locate the article!
The point being that technology advances will not be controlled by the US or anybody else and to attempt to do so shows the desperation evident.
Posted by: psychohistorian | Feb 12 2025 17:51 utc | 40
Posted by: Tom Pfotzer | Feb 12 2025 17:00 utc | 25
You're essentially right.
The first thing is to distinguish AI from Large Language Models (LLM). LLM are a subset of AI. Your examples (the chemical compound, optimizing a factory) are non-LLM AI.
Deepseek is an LLM.
non-LLM AI already are extremely useful, and China is leading in the field:
https://asiatimes.com/2025/02/will-deepseek-deep-six-the-us-economy/
Here Goldman presents 7 areas of competition in AI between the USA and China and argues that China leads in most of them.
Back to the general conceptualization behind this new industry.
All around us can be reduced to numbers (data) and mathematical relations. ALL. This writing, your reading, the movement of the air, your wife at home waiting for you (if you are lucky), your thoughts, ALL is numbers and mathematical relations.
AI collects all the numbers it can (big data) regarding a specific process and replaces the mathematical relations with statistical approximations, which means that non-linear mechanistic formulas are replaced with statistical linear approximations connected by neural networks. Since the statistical linear approximations are approximations, they can get better by learning (machine learning) with training data. Once the AI is well trained, it will output very fast very useful results, and will continue 'learning' as more data accrue.
This is current AI.
The CEO of Deepseek wants to develop AGI, the G is for 'general'.
This is when the statistical linear approximations can be phased out and truly non-linear mechanistic reasoning will be realized inside an AI=AGI.
That doesn't mean that AGI will be sentient, necessarily, but it will understand logic, cause and effect, and maybe that's all it needs to become sentient.
Posted by: Johan Kaspar | Feb 12 2025 17:54 utc | 41
they can get better by learning (machine learning) with training data.
Posted by: Johan Kaspar | Feb 12 2025 17:54 utc | 40
---
They cannot get "better". They can only produce the same answer more efficiently.
Posted by: too scents | Feb 12 2025 18:00 utc | 42
Posted by: CalDre | Feb 12 2025 17:06 utc | 28
Your opinion is contradicted by the dominance of open source statistical software.
In that industry, open source software relegated closed source commercial software to small niches.
And the open source software that became vastly dominant progresses and grows continually with numerous updates and a core group of maintainers and developers that formed a foundation.
The CEO of Deepseek is doing to AI what Ross Ihaka and Robert Gentleman did to statistical software.
Posted by: Johan Kaspar | Feb 12 2025 18:01 utc | 43
How many of us here use Linux or any other non-Apple, non-Microsoft O/S for our devices?Posted by: Fool Me Twice | Feb 12 2025 16:37 utc | 13
More than you think.
Posted by: ChatNPC | Feb 12 2025 18:04 utc | 44
That doesn't mean that AGI will be sentient, necessarily, but it will understand logic, cause and effect, and maybe that's all it needs to become sentient.
Posted by: Johan Kaspar | Feb 12 2025 17:54 utc | 40
Nah. With no body to incarcerate, nor soul to shame it can never be more sentient than the average corporation.
Maybe, one day, it might reach the sentience of an earthworm, but under US-based tech monopoly expect nothing other than long term, rent-bound stagnation.
See the iPhone/MS Windows/F-35 for a model of how that goes.
Posted by: ChatNPC | Feb 12 2025 18:12 utc | 45
Seems he's stuck making buggy whips for billions while the rest of the world is ready to move on. A dog barks in the night and the caravan moves on...
Posted by: Ledovik1 | Feb 12 2025 18:12 utc | 46
@ Posted by: Tom Pfotzer | Feb 12 2025 17:00 utc | 25
Apologies, I should have been clearer. I was referring to the specific AI products referenced in b's article and the current public debate. OpenAI, DeepSeek and the like are products designed to operate in the information space. To the best of my knowledge, they have no ability to conduct the types of original scientific research or industrial product development that you outlined.
Those types of scientific/engineering niche AI products are unlikely to ever be made available for general public use without tight controls and/or very high licensing fees.
Posted by: Fool Me Twice | Feb 12 2025 18:13 utc | 47
Don't forget that your 14 yr old knows more about your phone than you do... Idjits...
Posted by: Ledovik1 | Feb 12 2025 18:14 utc | 48
There are many examples of the best technology NOT prevailing.
That appears to be the signaling of the Vance AI conference speech.
Posted by: jopalolive | Feb 12 2025 18:17 utc | 49
The hypocrisy of the U.S. government is something to behold. They love to point their finger at everyone else, and cast the first stone, but never bring up their rotted past. Even ones that occurred on US soil, like the Trail of Tears, or the Bonus Army Massacre of 1932 or the Kent State shootings. And, in my opinion, the events of J6 could be looked at as the U.S.’s own Tiananmen Square event.
Posted by: Jose Garcia | Feb 12 2025 18:18 utc | 50
How many of us here use Linux or any other non-Apple, non-Microsoft O/S for our devices? Posted by: Fool Me Twice | Feb 12 2025 16:37 utc | 13More than you think.
Posted by: ChatNPC | Feb 12 2025 18:04 utc | 43
Lots of rebellious types here, perhaps more so in word than in deed, still I guesstimate <5%.
Perhaps Bernard can access his Google Analytics data and give us the real number.
As for the overall USA numbers...
In the US Analytics site, which summarizes DAP's data, you will find desktop Linux, as usual, hanging out in "other" at 1.5 percent. Windows, as always, is on top with 45.9 percent, followed by Apple iOS, at 25.5 percent, Android at 18.6 percent, and macOS at 8.5 percent.
https://www.zdnet.com/article/no-the-linux-desktop-hasnt-jumped-in-popularity/
Posted by: Fool Me Twice | Feb 12 2025 18:31 utc | 51
too scents @37
I suggest you look at the 10-K filings of the big players to get an idea of the juice.Only occurrence of the word "bribe" in their latest 10-K is assuring compliance with anti-bribery laws. What keyword should I be looking for? "Juice" doesn't appear. There is mention of referral fees to partners (which is generally third party consultants who get their clients to install the software, to quote, "earned upon the initial acquisition of a contract or subsequent upsell") and sales commissions, but obviously those fees are far less than the actual software licensing fees and, moreover, the money doesn't go to the client who is licensing the software (and instead could get open source software for "free", but chooses not to). So yeah, sellers and resellers get paid (as they do with any other product, too, including open source software).You could start with CrowdSrike.
Posted by: CalDre | Feb 12 2025 18:38 utc | 52
Posted by: too scents | Feb 12 2025 18:00 utc | 41
############
More data from more sources could arguably make it "better".
The AI can only answer what it "knows". The answers will likely change with different input.
Posted by: LoveDonbass | Feb 12 2025 18:50 utc | 53
Vance appeared at an Artificial Intelligence summit in Paris to push for the exclusive adoption of closed AI models which the U.S. attempts to monopolize...
Posted by b on February 12, 2025 at 15:59 UTC | Permalink
This is nothing new. Marc Andreessen was talking in an interview about a meeting with BIDEN:
https://www.youtube.com/watch?v=rulGP9cqS4w
I've tried three times now to post this with a transcript but it's not showing up so I'll keep out the transcript, let's see if this post shows up...
Posted by: Zet | Feb 12 2025 18:52 utc | 54
Johan Kaspar @42
Your opinion is contradicted by the dominance of open source statistical software. In that industry, open source software relegated closed source commercial software to small niches.Which open source software are you referring to? JASP, for example, is supported by the University of Amsterdam - i.e., government supported, which is exactly what I said was required for open source software to be successful (there are of course a few exceptions, not denying that, like the Linux kernel).
And the open source software that became vastly dominant progresses and grows continually with numerous updates and a core group of maintainers and developers that formed a foundation.A lot of the foundations are supported by government. But sure, there are some other ways to pay open source developers, sometimes through individual contributions, e.g. OpenAI at its inception, sometimes through mass (and government) contributions, like Libreoffice, etc., just like there are other ways to fund roads, but in general roads are built and maintained by government, due to the "free rider" problem.
The CEO of Deepseek is doing to AI what Ross Ihaka and Robert Gentleman did to statistical software.We really don't know exactly how Deep Seek is funded, allegedly by a hedge fund that was using AI for trading and the CEO decided to use those funds for a separate project. My guess it is paid for by the Chinese government since they allocated a ton of money for AI R&D. Which, if true, once again proves my general rule.
Posted by: CalDre | Feb 12 2025 18:53 utc | 55
I am just starting to read a related story from The Cradle
https://thecradle.co/articles/the-great-ai-game-us-china-vie-for-west-asian-cash
Posted by: psychohistorian | Feb 12 2025 18:53 utc | 56
Ok, posting the transcript doesn't work.
Listen to it, it's only a minute but you'll be shocked: the Biden admin already talked about restricting the whole AI ecosystem to two US companies. listen for yourself...
Posted by: Zet | Feb 12 2025 18:54 utc | 57
Posted by: Fool Me Twice | Feb 12 2025 18:31 utc | 50
The numbers you quote are from a market that is anything but free.
There is a significant hurdle for Linux to overcome against pre-installed desktop and phone OS's (as well as workplace coercion to use commercial product), plus a moderate amount of motivation and tech savviness required. Given that 1.5% is rather impressive.
Posted by: ChatNPC | Feb 12 2025 18:54 utc | 58
There is no such thing as "Artificial Intelligence". There is a desire to control and rule over people with technology, aka Technocracy.
Posted by: KOB | Feb 12 2025 19:02 utc | 59
Posted by: Zet | Feb 12 2025 18:52 utc | 53
Endorsing Trump didn't work out so well though.
The naivety of Tech Bros is legend.
Posted by: ChatNPC | Feb 12 2025 19:04 utc | 60
What is also interesting is that von der Leyen yesterday presented an EU initiative with 200 billion, which is supposed to advance the EU, and behind it is General Catalyst, one of the largest US investment companies, contributing 150 of the 200 billion...
Well, at least they quickly bought La Famiglia (a German investment corporation) from Princess von Fürstenberg shortly beforehand to at least look a bit European.
If you want to know who this princess is, it's the deep German oligarchy: "When the Princess calls to her palace, everyone comes. In early summer, Jeannette zu Fürstenberg (42) usually invites people to a royal tête-à-tête. Top executives, the middle classes and the tech scene then make a pilgrimage to Heiligenberg Castle on Lake Constance, which has been in the family since 1535. ... The princess even celebrated her 40th birthday in the glamorous Villa Medici in Rome together with Silicon Valley greats such as Palantir CEO Alex Karp (56)."
So you know where the journey is heading and as von der Leyen said: this is the biggest public-private partnership on the continent. Exactly what was decided at the G7 summit last year. In the end, there will be US AI in all areas: from healthcare to administration and education to the fight against crime.
The pieces of the puzzle are slowly falling in line.
Posted by: Zet | Feb 12 2025 19:08 utc | 61
Fool Me Twice @13: "How many of us here use Linux or any other non-Apple, non-Microsoft O/S for our devices?"
More than you seem to imagine. Desktops are going the way of the dinosaur. Many Gen Z types cannot even operate a mouse. Mobile devices are not just the future, but rather the present. Most people in the world use a mobile device, and most of those mobile devices run Linux (Android is just a flavor of Linux). Microsoft's efforts to grab market share on mobile devices was stillborn because Windoze sucks ass and was developed with the philosophy "Performance is a hardware problem".
Posted by: William Gruff | Feb 12 2025 19:10 utc | 62
Maybe the plan is to let only the rich get the good search results?
There is already a EU cloud with 42 countries that are the EU plus associates and the US plus 5 eyes basically. By outcasting anything Russian or Chinese they want something "wasp sourced only".
Posted by: Tom | Feb 12 2025 19:11 utc | 63
Below is a quote of the closing paragraph of the article from The Cradle I linked to in my comment #55
West Asia: The battleground for AI influenceAs AI competition intensifies, West Asia is no longer just an emerging market – it is a strategic theater in the tech war between Washington and Beijing. China views the region as an extension of its Digital Silk Road, aiming to expand its technological footprint through cost-effective AI solutions.
The US, on the other hand, is deepening its AI partnerships with Persian Gulf states, trying to ensure that AI infrastructure aligns with western standards. The battle over AI in West Asia transcends mere technological rivalry; it is a contest for economic and geopolitical dominance.
With Persian Gulf states positioned as kingmakers in this struggle, their decisions in the coming years could redefine the balance of power in the AI era. The US–China AI war is no longer just a two-player game – West Asia is now firmly in the mix, and its role in shaping the future of AI is only growing.
I choke on the "aligns with western standards" trope that so many use to not refer to the private financial jackboot that the West is trying to maintain....but otherwise worth reading article
LLM type systems do have great potential in many areas but this delusion that any computer program can do more than what it is told to do fits our hubristic tendency just like all religions....many refuse to accept how ignorant we are....below is part of a comment I made last night in the open thread
To all the science barflies I want to share today's hubris killer for me, Astronomy Picture of the Day
https://apod.nasa.gov/apod/astropix.html
Today the subject is Asteroid Bennu Holds the Building Blocks of Life
The mind bending part
What can a space rock tell us about life on Earth? NASA's OSIRIS-REx spacecraft made a careful approach to the near-Earth asteroid 101955 Bennu in October of 2020 to collect surface samples. In September 2023, the robotic spaceship returned these samples to Earth. A recent analysis has shown, surprisingly, that the samples contained 14 out of the 20 known amino acids that are the essential building blocks of life. The presence of the amino acids re-introduces a big question: Could life have originated in space? However, the protein building blocks themselves held another surprise -- they contained an even mixture of left-handed and right-handed amino acids -- in contrast to our Earth which only has left-handed ones. This raises another big question: Why does life on Earth have only left-handed amino acids?
So, let me know when AI can tell us why life on Earth have only left-handed amino acids?
Posted by: psychohistorian | Feb 12 2025 19:12 utc | 64
The Moscow Exchange Index flies into space.
Gazprom added 10%
Posted by: HughG | Feb 12 2025 19:16 utc | 65
developed with the philosophy "Performance is a hardware problem".
Posted by: William Gruff | Feb 12 2025 19:10 utc | 61
---
AI fin-tech crowd: "Let's run that up the flagpole and see who salutes".
Posted by: too scents | Feb 12 2025 19:16 utc | 66
There is already a EU cloud with 42 countries that are the EU plus associates and the US plus 5 eyes basically. By outcasting anything Russian or Chinese they want something "wasp sourced only".
Posted by: Tom | Feb 12 2025 19:11 utc | 62
As one of the next steps they will introduce electronic IDs which you'll have to use to prove your age when watching adult content on the Internet. They will use kids safety and these adult content age checks as a smokescreen to get this rolling, see the decision about limiting social media access etc.
Besides deanonymizing, tracking & control this can then later also be used to restrict access to certain services; the ultimate silicon curtain if you can only log into Google with an electronic ID from one of the Western nations.
Posted by: Zet | Feb 12 2025 19:18 utc | 67
let me know when AI can tell us why life on Earth have only left-handed amino acids?
Posted by: psychohistorian | Feb 12 2025 19:12 utc | 63
---
The symmetry mystery is so obvious you'd have to be biased to miss it.
Posted by: too scents | Feb 12 2025 19:21 utc | 68
Ukrainian bonds which are trading a 42 cents on the dollar or the hryvnia or whatever has jumped 10% on the phone call between Putin and trump as peace seems more likely.
Bonds, real money, are better indicators of future events than poster posting continuously whom , obviously, have no real job
The symmetry mystery is so obvious you'd have to be biased to miss it.
Posted by: too scents | February 12, 2025 at 19:21
---
Adding.
How would you even program a left-handed or right handed AI? We are clueless.
Posted by: too scents | Feb 12 2025 19:23 utc | 70
@Johan Kaspar, FoolMeTwice and the others that commented on my post: thank you. Especially Johan. That was excellent, and I'll read closely the link you provided.
=== Regarding open source.
I see Open Source as a pre-eminent, immensely powerful equalizer between the powerful and the weak(er). Open Source harnesses the powers of the many, and is an almost-unprecedented example of highly coordinated and yet not-centrally-controlled emergent behavior.
The Linux community - not just the OS, but also the perfectly immense library of development tools and applications, many of which are or are close to being best-of-breed - has already vastly altered and expanded the trajectory and maneuvering room of the "little people" and restricted to some great degree the trajectory and maneuvering room of the "strong".
For example, see how Microsoft is integrating Linux into its broad set of offerings. Note that a good bit of the "cloud" uses Linux as the platform upon which their myriad offerings are constructed. In early 2000s, Microsoft CEO Steve Ballmer infamously compared Linux licensing to “a cancer that attaches itself in an intellectual property sense to everything it touches”.
Well, what a difference a few years makes.
Linux is a Commons which was built by little people (and some Bigs) and is available to all without license fees (mostly). It has an extremely well-entrenched and highly developed "ecosystem" in the same sense that DeepSeek wants to build such a Commons. DeepSeek has picked out the right strategy to vanquish the Bigs.
Even if, somehow, the West prevents DeepSeek and its philosophical cohorts from getting entrenched in the West, the sheer numbers of people from India, China, Iran, Russia and so forth that will contribute to that Commons will easily out-run the rate, direction, and depth of development that the West can field.
The West will lose that footrace, and maybe the West has already lost it.
One thing that jumps out at me is the sheer volume of investment the West has (is currently, wants to continue to) make in AI-supporting data centers. It's astonishing. The West is (seriously!) considering building small-scale nuke plants _just to service these data centers_.
Think about how heavily invested the West is in brute-force AI. That's one heck of a lot of infrastructure build-out we've committed ourselves to, and look at how vehemently we're defending that strategy. Speaks volumes.
And so it becomes more obvious why Mr. Vance said what he said, when he said it, and where he said it.
This U.S.-led AI commitment may well be - I don't know yet, but I wonder - may well be another own-goal; another case of letting hubris and access to very deep pools of money dictate the U.S. national strategy.
We used similar technique and perspective to commit the NeoCon-Iraq, NeoCon-Afghanistan, NeoCon-Ukraine gambits. It would not surprise me to find out, shortly, that the AI project - given its scope, intensity, rapidity, and policy-lockstep across government and industry ... is the latest world-domination gambit. That now seems rather obvious, right?
Now consider the Chinese mentality. They don't do frontal warfare. They do infrastructure warfare. Massive manufacturing infrastructure investments. BRI. Alibaba. Container-ports. Ship-building. China Post. Electric cars. Domination of solar business. High speed rail. DeepSeek is another brick in the Infrastructure-War strategy.
It seems very likely to me that everyone one outside the West will go gang-busters to insure that they're part of the DeepSeek-type Commons-building. It's strategic for them, and Mr. Vance has made that point with a figurative mushroom-cloud.
Open source is a great equalizer. It leverages the strength of the many, and it works. Bigtime. Expect a lot more of it, and from all corners of the world.
Posted by: Tom Pfotzer | Feb 12 2025 19:28 utc | 71
William Gruff @61
Most people in the world use a mobile device, and most of those mobile devices run Linux (Android is just a flavor of Linux).Linux is just a kernel and a tiny part of the AndroidOS software stack. If you use Firefox on Android you are getting far more "open source" than by using the kernel. Linux kernel by itself is utterly useless. Everything on top of it - and even part of it, as many drivers - are proprietary and closed. That's why there aren't many AndroidOS forks. The most popular one is LineageOS, which isn't even a blip on the OS market share graphs (~ < 2 million installations; AndroidOS has ~ 12 billion installations).
But sure, lots of people use Linux directly and indirectly. I think the person you were responding to meant GNU/Linux, i.e., a Linux-based desktop like Ubuntu, Debian, Red Hat, CentOS, etc. That is also a tiny percentage, though far higher (ironically) than for smartphones.
Posted by: CalDre | Feb 12 2025 19:39 utc | 72
Linux is a Commons which was built by little people
Posted by: Tom Pfotzer | Feb 12 2025 19:28 utc | 70
---
Linux was a Commons.
Now it is just another FoundationCorporation.
https://www.linuxfoundation.org/about/members
Russian Linux kernel maintainers blockedTo ensure compliance, the Linux kernel will no longer allow Russian software developers to work on maintaining the codebase
28 Oct 2024 16:11
Last week, Greg Kroah-Hartman, the current maintainer of the stable branch of the open source Linux kernel, issued a message on the Linux code maintainers mail list announcing that it was removing some developers due to compliance requirements.
“They can come back in the future if sufficient documentation is provided,” he wrote in the message sent to the Linux patch list of recipients who help maintain the kernel code.
This mailing list includes Linus Torvalds, the developer of the original Linux kernel. In a post to the same Linux patch maintainers mailing list, Torvalds spoke about his concerns that there were lots of Russian trolls who could potentially infiltrate the Linux kernel. “It’s entirely clear why the change was done. It’s not getting reverted, and using multiple random anonymous accounts to try to ‘grass root’ it by Russian troll factories isn’t going to change anything,” wrote Torvalds.
However, some argue that the decision to remove the Russian developers was not transparent.
...
https://www.computerweekly.com/news/366614656/Russian-Linux-kernels-maintainers-blocked
I've been using Linux since 0.9. There has been a sea change since the rise of the post Global Crossing tech-bros.
Thank goodness for Richard Stallman.
Posted by: too scents | Feb 12 2025 19:40 utc | 73
How many of us here use Linux or any other non-Apple, non-Microsoft O/S for our devices?
How many people use Android phones. Then they are using Linux.
How many Data Analysts or people performing Data Analytics use Python. Then they are using a programming language that uses a GPL license.
As others have pointed out, on backend servers and embedded in Operating Systems is Linux. Heck, even Microsoft makes a version of SQL Server that runs on Linux.
The point is Linux and Open Source are used more widely then you know. Linux is used where the REAL money is to be made, backend servers and mobile devices.
Posted by: JoeSixPack | Feb 12 2025 19:49 utc | 74
too scents @72
Now [Linux] is just another Corporation. ... Russian Linux kernel maintainers blocked
This seems to be related to US sanctions against Russians entities, which are the "compliance requirements" noted in the announcement. "They can come back in the future if sufficient documentation is provided" seems to imply they can return if they can demonstrate they are not sanctioned persons or working for sanctioned persons. See https://sfconservancy.org/blog/2024/dec/12/linux-banned-russian-contributors-do-i-need-to/
Obviously Linus' responses were unhelpful - as has long been apparent, he's an utter idiot outside of kernel issues, and he clearly hates Russia and Russians in line with current Finnish (which he is) hateful thinking. But he's not a "corporation" - he's an "individual". Shocking that an individual can be an idiot and do stupid things, huh?
Posted by: CalDre | Feb 12 2025 19:56 utc | 75
Shocking that an individual can be an idiot and do stupid things, huh?
Posted by: CalDre | Feb 12 2025 19:56 utc | 74
---
Torvalds is in the Linux Foundation's pocket.
Posted by: too scents | Feb 12 2025 19:59 utc | 76
Torvalds is in the Linux Foundation's pocket.Well, he's paid by them. Most of the funding comes from its platinum members: Ericsson, Fujitsu, Hitachi, Huawei, Intel, Meta, Microsoft, NEC, Oracle, Qualcomm, IBM/Redhat and Samsung. Each of them gets one board seat, for 12 out of the 21 total (3 are "at-large", 3 are elected by the gold members and the rest are Linux Foundation officers).
So sure he is paid by Linux Foundation but it's not clear they could actually depose him as project leader, that would interesting if it were attempted. But in any case the LF itself is not some monolithic entity.
Posted by: CalDre | Feb 12 2025 20:16 utc | 77
But in any case the LF itself is not some monolithic entity.
Posted by: CalDre | Feb 12 2025 20:16 utc | 76
---
Imagine the Linux Foundation allowing strong cryptography into the mainline kernel back in the day.
A new version of the international crypto patch for 2.1 has been released by Alexander Kjeldaas. The purpose of this patch is to gather together all of the various cryptographic patches that exist for Linux into one convenient place. Due to silly laws in a few countries, this stuff will probably never be part of the official kernel, but is certainly useful anyway. See the announcement for information on what's included and how to get it.
Note how the patches were made available internationally to skirt export restrictions.
That would never fly today.
Posted by: too scents | Feb 12 2025 20:24 utc | 78
Excellent summary here:
DARK GOTHIC MAGA: How Tech Billionaires Plan to Destroy America
https://www.youtube.com/watch?v=5RpPTRcz1no
Posted by: Zet | Feb 12 2025 20:34 utc | 79
too scents @77
That would never fly today.I don't know, have a look-see at this. Which standard crypto technologies are not available for the recent kernels?
And in any event any third party can make patches, I've often patched my kernel, isn't that hard, certainly not for Linux OS creators/distributors.
Posted by: CalDre | Feb 12 2025 20:37 utc | 80
@Scottindallas #4, 8
Sorry, bud, but you are wrong in your assertion that the LLMs cannot process humor or satire or whatever. The one thing they can do, is handle language well via their point clouds around each word and dictionaries of word usages.
The real problems of LLMs are that they really don't do anything super well.
For example: you can use LLMs like the Musk boys are - to organize and query large masses of data. But this is a limited productivity add and in reality, a reduction in jobs. Instead of say, a cascade of layers where information is gathered, organized, structured, queried and the analyzed - something which required some people but less than most think, now you can do the same job with perhaps 1/2 or 1/3 of the middle and front end types...assuming the data is good. If the data is not good, then the savings is negative.
But there just are not many jobs where you need this type of work done.
I doubt there are 100K people doing this type of work in the US now - it is probably closer to 10K. Firing half to 2/3rds of them is not going to change anything significant and is not going to be worth investing $500B+ over.
Posted by: c1ue | Feb 12 2025 20:39 utc | 81
@Jeremy Rhymings-Lang #9
The most energy efficient solution will not win out if massive energy use, ginormous capital cost and $100 billion single model training costs are seen as a feature - a means towards monopoly/oligopoly - by the US tech companies.
That's exactly what Ed Zitron writes about.
Posted by: c1ue | Feb 12 2025 20:41 utc | 82
I've often patched my kernel, isn't that hard
Posted by: CalDre | Feb 12 2025 20:37 utc | 79
---
Self signing kernels so that they will secure boot is a PITA!
Posted by: too scents | Feb 12 2025 20:44 utc | 83
@Tom Pfotzer #25
Totally wrong.
LLMs don't understand jack shit.
What they do is create probability clouds around each word and a notation of the ways that word is used (verb, noun etc). A convolution of multiple words traverses the resulting multiple probability cloud induced branching paths, constrained by the random exercise of each probability cloud along its constrained paths (word type), and is then further pruned by the "training" to remove known bad paths.
This is precisely why LLMs output garbage regularly: there are enormous numbers of wrong paths which are not covered by training data.
A small subset of these "wrong" paths are held up as "intelligence" when all it is, is cherry picking garbage.
The scam put forth by OpenAI/Anthropic, AS WELL AS, Openseek, is that there is any way to fix this problem. There is not. The solution set is way too large - it makes the millions monkeys/million years/million typewriters look like nothing, but the scammers are betting that they can scrape all of the internet such that the long tail of junk paths will be relatively rare.
The above is also why LLMs will make shit up. There are perfectly grammatically correct but factually wrong ways by which the probability cloud traverses the possible paths such that LLMs will pull imaginary legal references, published books, etc etc literally out of its ass.
Posted by: c1ue | Feb 12 2025 20:53 utc | 84
too scents @82
Self signing kernels so that they will secure boot is a PITA!Not that hard, there are tools for that. But in any event UEFI is hardly foolproof, since you can (even if a PITA for you, not for a hacking team) self-sign kernels, and in fact the "big boys" (like NSA) can install a compromised (and signed) kernel on your device. It's really quite limited protection. If you want security, don't connect to the internet and lock your computer in a vault. That's what the "big boys" do for their secure environments.
Posted by: CalDre | Feb 12 2025 20:55 utc | 85
@CalDre #28
LOL open source.
Open source was a thing when there were relatively few programmers and relatively little code.
Paul Vixie was one of the 2 people famous for pushing open source in the 2000 time frame, but he recanted.
He recanted because when he made that statement, there were millions of lines of open source code and the open source community was reasonably able to review the code body.
By 2019, when he recanted, open source code had ballooned to tens of billions of lines of code while the community was about the same size. There is no way in hell that anyone other than the author looked at 95%+ of the open source code outside of a small core of base functionality.
The notion of government support for open source is even more stupid. The entire point of open source is that the people who use the code, will put in the effort to make better code and/or review the code the use.
A government paid bunch of open source maintainers would convert open source into the equivalent of the DMV - the driver's license bureaus which are famous for working, more or less, but being incredibly slow and inefficient.
Hard Pass.
Posted by: c1ue | Feb 12 2025 20:58 utc | 86
even if a PITA for you
Posted by: CalDre | Feb 12 2025 20:55 utc | 84
---
When I say PITA I mean that an additional obstacle has been introduced in the goal of running my software that provides absolutely no benefit.
Of course I can dis-assemble and re-flash the BIOS/UEFI firmware to my hearts content. But why should I have to?
Why should Microsoft Corp. hold the root key to my device?
Posted by: too scents | Feb 12 2025 21:02 utc | 87
@Johan Kaspar #40
You are correct in that there are many forms of "AI" that are not LLM.
You are not correct that these are 'intelligent' in any way. All they are, are various forms of models trained by data.
All models can be useful but no model is reliable.
A simple example: machine vision is one of the earliest "self training" type models used. This is an area which has been worked on for literally decades.
And yet machine vision is still shit compared to what every animal on earth with eyes has.
It is not hard to understand why. Biological vision is a capability that has been honed in the harshest environments, with incredible diversity, for literally millions of years with absolutely clear parameters: work or die.
The scam from the "self training" believers is that human data can match the naturally evolved vision via "smart" data.
The proof is in the pudding, and there is no machine vision pudding still.
Posted by: c1ue | Feb 12 2025 21:05 utc | 88
@Tom Pfotzer #70
LOL open source as an equalizer.
You do realize that the vast, vast majority of open source contributors work for big software companies, right?
Posted by: c1ue | Feb 12 2025 21:07 utc | 89
@JoeSixPack #73
Thank you for the laugh of the day!
Seeing anyone call out Android - the darling of Google - as a champion of open source is the funniest shit I have read in a long while.
Next you're going to try and tell me that Bill Gates and Microsoft built DOS and Windows because they are altruists.
Posted by: c1ue | Feb 12 2025 21:09 utc | 90
There is no way in hell that anyone other than the author looked at 95%+ of the open source code outside of a small core of base functionality.Sure, and from the get it was never guaranteed anyone would. The thing is everyone can (even though that assumes that the distributed binaries are compiled from the distributed source code, which itself is quite hard to prove, but you can avoid that issue by compiling yourself). On the other hand, you have no chance of looking at proprietary source code (except for some really large licensees who demand source code copies), but people use it anyway.
The notion of government support for open source is even more stupid. The entire point of open source is that the people who use the code, will put in the effort to make better code and/or review the code the use.That comment is incredibly stupid, since government funding for developers does not stop that mechanism, and in any event that is not the primary argument for open source. Though of course the government could also fund alpha/beta testers, making it even more consequential. Not to mention that multiple governments can fund development / testing of the same software, allowing this global collaboration which is essentially what you are stupidly mocking.
A government paid bunch of open source maintainers would convert open source into the equivalent of the DMV - the driver's license bureaus which are famous for working, more or less, but being incredibly slow and inefficient.You are of course correct, government funding is why we don't have any public roads, airports, canals, or utilities. Sigh. But sure, if you do it stupidly you can waste a bunch of money. On the other hand, Microsoft Office, which has been around for many decades, still makes $45 billion per year, surely there's no waste or gouging in that, eh?
Posted by: CalDre | Feb 12 2025 21:10 utc | 91
@CalDre #90
It is quite clear you don't know what the fuck you are talking about.
If no one but the author looks at the code - the probability of even just typos getting in there is very high, much less the other outcomes ranging from incompetence to deliberate malice.
In a big corporation, they actually do a better job of reviewing the code - especially these days with the focus on security.
Posted by: c1ue | Feb 12 2025 21:13 utc | 92
too scents @86
When I say PITA I mean that an additional obstacle has been introduced in the goal of running my software that provides absolutely no benefit.The benefit is you are getting the cryptography you think you otherwise may not get due to the Linux Foundation being a tool of ... something ..., or even worse - shudder! - is a ... wait for it ... corporation! ... for free. (I note you have not identified any crypto you cannot so get, but I digress.) And you can easily disable secure boot, which basically helps nobody anyway.
Why should Microsoft Corp. hold the root key to my device?Complain to Dell and Lenovo. But my hypothesis is: they produce 90% of the operating systems running on computers with UEFI. If your OS had the same marked share, I'm sure you'd get a key too. Heck even smaller OSes get a key. But obviously the whole point is defeated if everyone has a key.
Posted by: CalDre | Feb 12 2025 21:20 utc | 93
Trump, Vance et al. should make friends with China. Not everyone is either a slave or an enemy, Trump!
Posted by: lester | Feb 12 2025 21:20 utc | 94
The scam from the "self training" believers is that human data can match the naturally evolved vision via "smart" data.
Posted by: c1ue | Feb 12 2025 21:05 utc | 87
Just a question: in which camp are you exactly?
a) this whole AI is bullshit, doesn't have any real valuable use case whatsoever
b) this whole AI has great potential but it certainly doesn't solve all the problems which currently are promised to get solved and the hype is totally nuts
c) other, please explain
Where AI is just the current code / buzzword around all the technologies regarding machine learning / intelligence simulation, from Narrow AI (the current state of things, LLMs, chatbots, generative AI), to General AI (we are not there yet, just some tiny steps, see recent developments around physical AI, world AI), to Theory of Mind AI or maybe even self-aware AI.
Posted by: Zet | Feb 12 2025 21:21 utc | 95
c1ue @91
It is quite clear you don't know what the fuck you are talking about.Oh, I love projections. Can I have some more please?
If no one but the author looks at the code - the probability of even just typos getting in there is very high, much less the other outcomes ranging from incompetence to deliberate malice.You'd be amazed how much commercial code is not reviewed by anyone but the author. But in any event, how do you know how many people review each line of, say, Google Maps or 7Zip code?
In a big corporation, they actually do a better job of reviewing the code - especially these days with the focus on security.You have a study on that, or is that just your omniscience blessing us mortal idiots altruistically?
Posted by: CalDre | Feb 12 2025 21:23 utc | 96
c1ue @91
Also, in terms of malice, how do you know these "well-reviewed" corporate lines of code don't insert backdoors? Since, in fact, there is lots of evidence that ... they do. But I guess you can't review their code to check, eh? Funny, that. Maybe your palpable omniscience has cleared it all?
Posted by: CalDre | Feb 12 2025 21:26 utc | 97
A lot of people talk about Deepseek and Tianamen Square. How do ChatGPT & Co perform with queries on, say, Gaza, the war in Ukraine, Taiwan, the Iraq Wars, january 6, Twittergate, ...?
Posted by: Marvin | Feb 12 2025 21:30 utc | 98
Open source was a thing when there were relatively few programmers
Posted by: c1ue | Feb 12 2025 20:58 utc | 85
---
There are always relatively few programmers. Just like there are relatively few pianists.
Anything that requires skill will have relatively few practitioners. The more skill required the fewer masters there will be.
It is always the same in human endeavors. One or two outstanding individuals carry progress forward.
Now, outstanding individuals has become a social paradox for China to solve. As luck would have it even geniuses' deficits are made whole by the fabric of society.
Easy. Happy.
Posted by: too scents | Feb 12 2025 21:32 utc | 99
If American technology was better than Chinese technology, the US government wouldn't feel the need to force them to use American by putting a gun to their head. Or to use sanctions.
Posted by: Marvin | Feb 12 2025 21:34 utc | 100
The comments to this entry are closed.
Pete Hegseth just told Europe and Ukraine to “GFY” by announcing that Europe and Ukraine were no longer top security concerns for the US. I’d say the odds are even that the US pulls out of NATO or diminishes its support to the point it collapses within a couple of years.
Posted by: CullenBaker | Feb 12 2025 16:09 utc | 1