Forgot your password?
typodupeerror

Comment Re:Open Source Wins Again (Score 4, Insightful) 55

To be fair, you need at least 256GB of RAM just to run the 2-bit version of this model. Most people aren't going to be able to do that at home.

But yeah, the Chinese government is willing to throw lots of money at building AI models and giving them away, so Western companies are screwed.

Another way of looking at this if western companies are screwed.. hardware prices return to planet earth where more people are able to run this stuff at home. Three years ago the cost of 512GB DDR5 was less than the cost of a single 4090 GPU today.

Comment Re:It doesn't (Score 2) 55

I like free and local LLM, but they do not match the Frontier models of Antropic, OpenAI or Google yet.
The take here is either uninformed, sensationalist or fear mongering (possibly to ban Chinese LLM). GLM is good, but nowhere near Mythos/Fable or even latest Opus.

What is the objective basis of your statement? Do you reject benchmark results? If so what have you used in their place to make such a determination?

Comment Re:Next bubble (Score 1) 73

You cannot pick a handful of data points and then choose the type of curve to fit them, and that's especially a problem when you have only a few data points. You are using literally three data point. For any tech thing, if you ended up with a curve that shows negative progress, that should be a problem with your model. This is essentially the same thing the Trump people tried to do when they tried to do a cubic fit for covid deaths to argue that deaths would soon drop to zero https://www.vox.com/2020/5/8/2.... But since you don't see where the exponential estimate is coming from, you could take the step of clicking through the link I gave which discussed it. But if you want here are other sources. For example, https://physics.aps.org/articl... discusses how decoherence times have gone up at an exponential rate, increasing by roughly a factor of 10 every 3 years. You are correct that successful use of Shor's algorithm has not gone up but that also shouldn't be surprising. Shor's algorithm has a pretty big jump in the number of logical qubits needed when you increase from very small n to medium sized n. Using those two data points to conclude anything about what is going on right now isn't useful. Once it does hit even n around 105 or so, we should expect then quick improvement from there.

What data points do you have to support exponential growth of logical qubit count with time?

Comment Re: "Powerful" quantum computer (Score 1) 73

No need to feel ashamed. If you want, I can teach you the difference between the two. I can also provide English language services to assist you as well.

Please yes, I would love an explanation how your reference is responsive to "Wake me up when they create one with even 100 reliable logical qubits." given the experimental device you referenced did no such thing.

Comment Re:Let me guess (Score 1) 161

another DEI/SJW film, and the people involved are baffled about why it failed?

My initial knee-jerk reaction is that it isn't woke enough which is why woke outfits like the bulwark are calling it boring. These are the same people who called the everything everywhere random AI slop movie "profoundly moving" and "great".

Keep in mind I have zero clue what is actually in this movie and most probably will never see it.

Comment Re:Even so... (Score 2) 114

All of it. It doesn't require specific motherboards

Of course it does. TPM keys are stored in HARDWARE onboard a specific motherboard. This creates unnecessary and unwanted dependency on specific hardware.

TPMs are also built in to all CPUs from the last decade, and the firmware for TPMs are rarely touched.

On my PC the TPM is wiped whenever a BIOS update is installed. This behavior is extremely common.

Comment This is an interesting topic, at least to me. (Score 1) 2

I have been stress-testing AIs with increasingly complex projects for some time. The Chinese AIs struggle, but actually do a FAR better job of handling massively complex tasks than Grok, and Gemini just rolls over and whimpers at anything above a very low level of complexity.

What I've found is that the Chinese AIs tend to be sycophant but do "understand" complex projects properly in that you can ask specific technical questions and the answers will be generally very accurate. Any sort of critical analysis is beyond them, though. (Ether that, or I'm a mega-genius. Which....doesn't sound terribly likely.)

Of the "Top AIs", ChatGPT is good on basics but is incapable of any kind of detailed generation. Claude is brilliant at detailed generation, but overloads with anything but a tiny data set.

I've been putting up the projects on Gitlab for a while, so anyone who wants to see an AI break down and cry in despair is able to do so.

The secret tools don't bother me - they'll have long understood how to use Big Data and Analysis of Competing Hypotheses. AI isn't going to find out any more than combinations of those tools will, because that's basically all AI is - a Big Data classification system.

Submission + - The MOST artificial intelligence is Chinese? (linkedin.com) 2

shanen writes: Pardon my clickbait and quasi-joke Title suggestion, but the topic has been on my mind for a while. I have not been pursuing the research topic seriously, though I did take several close looks at DeepSeek when it was the center of hoopla and have sometimes benchmarked against it since then. But this summary of new Chinese AI was just pushed at me by the AI-empowered algorithms of LinkedIn and I'm wondering how seriously I should take it.

If we (non-Chinese?) were actually technically ahead of them (Chinese heathens?) then this would not be an issue. Unlike the computer security race we lost a few years ago? However the real concern is not with these public AI tools, but with the secret ones, both government and private... (Bond villain conspiracy theories, anyone?) But I don't think there is likely to be an outspoken and authentic expert from inside China also inside the (Slashdot) house.

Personal disclaimers needed? Lately most of my AI games of the non-fun type have involved Claude, but Gemini keeps sticking it's remarkably unintelligent nose into my business to the point where I've become much more tolerant of Bing than I used to be. More broadly, there used to be a time when I would have high confidence of seeing useful discussions on Slashdot with some known experts who were probably the real people to boot (in at least two senses of "real"), but these days Slashdot has also been infected with the lack-of-trust virus. Another terminal case? I can't say, but I'm no longer surprised when one of the oldtimers keels over. Bash.org had a great collection of jokes...

Comment Ok. (Score 4, Interesting) 81

So you're telling Claude something vague and washy, then Claude invents a prompt that might vaguely possibly be somehow related to what you want along with a drink that is almost but not entirely quite unlike tea. Claude then recurses through this until it has a Celtic knot so intricate that it has its own Hausdorff dimension. What burps out is a product that is completely useless and patented to the Sirius Cybernetics Corporation.

Comment Re:Time for Microsoft to do a Coca Cola (Score 1) 114

My favourite part about your comment is it shows that people will eventually come around to calling the current mistake "classic" and demand it. Your post would not have been out of place back in 2015 lamenting about the loss of Windows 7 and lambasting Windows 10.

Why would anyone give two shits about unnecessarily disruptive change when there is little to no commensurate value to show for it in return? This isn't the 90s. PCs and operating systems are a mature technology. For many there is more value in continuity.

These days the value proposition is often negative given Microsoft's malware oriented business model depends upon increasing aggression towards its own customers. Endless ads, spying, embarrassing UX regressions and unwanted dependencies rather than useful value.

Slashdot Top Deals

Biology is the only science in which multiplication means the same thing as division.

Working...