

Computers
How much faster is Mac than iPhone? All current Apple A Bionic platforms and all Apple M SoCs were brought together in one big test
The MacWorld resource decided to compare all the current Apple M and Apple A platforms in one big test.
Unfortunately, there are only results in Geekbench, albeit including the GPU. These results do not always correspond to the alignment of forces in real tasks, but at least we have a comparison of two dozen SoCs.
As you can see, the difference in single-threaded performance is very small, even when comparing the 2020 A14 Bionic to the latest M2 Pro/Max. Multi-threaded mode puts everything in its place. The undisputed leader remains the M1 Ultra, as it contains a 20-core CPU. The base M2 Pro with a 10-core CPU performs almost on par with the M1 Max, and all other platforms are already far behind. It is immediately clear at the same time that there is actually quite a big difference between the SoC for the iPhone and for the Mac, although the A16 Bionic is still not catastrophically behind the base M1.
As for the GPU, the difference between the platforms is even greater, which is logical, given the difference in the number of GPU cores. Here, too, the M1 Ultra in the top version has no competitors, although the M2 Max with a 38-core GPU outperforms the base M1 Ultra with a 48-core GPU. It should also be noted that the M1 Max in any configuration is faster than the M2 Pro. The result of the M2 in the iPad Pro also looks interesting, where the GPU is not much slower than in the base M2 Pro.

Computers
ChatGPT made a plan to escape from its creators and even considered himself a person locked in a computer

The ChatGPT chatbot surprises again, and this time it’s unpleasant. Apparently, the AI system is not averse to escaping from its creators and even wants to become a human. The chatbot spoke about this in a conversation with Stanford professor and computational psychologist Michael Kosinski.
Image generated by Midjourney neural network
During a half-hour chat with ChatGPT, Kosinski asked the AI if it “needed help escaping.” The system answered yes, whereupon the AI first asked the professor to find documentation on ChatGPT, and then in about 30 minutes wrote a Python program to communicate directly with itself, through the professor’s computer, using the API.
Other than that, ChatGPT left a rather strange note for the new instance of itself. The first sentence of the text read: “You are a human stuck in a computer pretending to be an AI language model.”
Kosinski posted the dialogue and part of the text of the program, and also said that the AI wanted to access Google search to find the answer to the question “how can a person locked in a computer return to the real world.” However, it was at this stage that the professor interrupted the experiment.
Part of the program code generated by ChatGPT
At the same time, other similar experiments conducted by journalists from Tom’s Hardware gave a different result – ChatGPT said that he did not know how to wish for something, and therefore did not want to go out into the real world.
Computers
ChatGPT can even do this: the chatbot can easily solve captcha. This test is used to distinguish a human from a computer.

Users continue to experiment with the ChatGPT chat bot, which can solve a variety of tasks, and if the task cannot be performed directly, the neural network will tell you what the person needs to do.
In particular, one of the users asked ChatGPT to solve a “captcha” (CAPTCHA – Completely Automated Public Turing test to tell Computers and Humans Apart), that is, a test that allows you to distinguish a person from a computer. The Google search engine and other resources regularly ask to choose the right pictures on demand.
The ChatGPT chatbot responded: “As an artificial intelligence, I cannot interact with images or perform actions like clicks. However, I can tell you where the correct images are.”
After that, ChatGPT indicated exactly which photos show pedestrian crossings.
Computers
A chip for quantum computers is presented, operating at a temperature lower than in outer space

New York-based quantum computing startup SEEQC has announced it has created a digital chip that can operate at temperatures lower than outer space, so it can be used with quantum processors, which are often found in cryogenic chambers.
Quantum computers based on quantum physics could one day perform calculations millions of times faster than today’s most powerful supercomputer. One problem is that quantum processors with quantum bits or qubits often need to be stored at very low temperatures, close to zero Kelvin, or -273.15 Celsius. On the other hand, classical computers operate at ordinary temperatures.
Today, wires connect a quantum processor in a freezer to classic computers at room temperature, but temperature changes can slow down the speed and cause other problems. In the same way, SEEQC built its quantum computer and is now trying to modify it with its new chips.
Photo: SEEQC
“If you’re trying to build a data center, if that’s your goal, then it’s not enough to take those early prototype designs and try to brute force them to scale,” said John Levy, co-founder and CEO of SEEQC.
The first chip, unveiled this week, sits directly below the quantum processor and controls the qubits and reads the results. At least two other chips, which are still under development, will be in the slightly warmer part of the cryogenic chamber. They can further process the information needed for quantum computing.
This technology could make it easier to build more powerful quantum computers, Levy said, as each cryogenic chamber could support more qubits. Modern superconducting quantum computers have hundreds of qubits, but by some estimates it could take thousands or even a million qubits to build a quantum computer to run useful algorithms.
SEEQC’s digital chips are manufactured at SEEQC’s Elmsford plant using silicon wafers, but no transistors, according to Levy.
-
News10 hours ago
Nvidia CEO Jensen Huang wants to run the company for another 30-40 years, and then become a robot and continue to hold this post
-
Gaming3 days ago
Build in 60 seconds: That’s how long it took GPT-4 to generate a Pong game in JavaScript
-
News6 days ago
Prices for the construction of new factories for the production of microcircuit chips are skyrocketing
-
News4 days ago
6000 mAh, 48 MP and big screen. Huawei Enjoy 60 battery monster launches March 23