Sensei

In every industrial sector, they aim to become a monopolist and make their customers dependent on their products, as this ensures reliable, annual profits—which are essential for the company’s operations and for covering the costs of all their highly expensive specialists. In every industry, the goal is to increase revenues and profits year after year. A manager who shows a decline will simply be dismissed as ineffective. So don't blame IT, blame capitalism. That's how it works. Why is there a new version of Android and a new version of iOS every single year? Because companies hire programmers who need to be paid—otherwise, they might leave for somewhere else. IT companies are prisoners of progress. Everyone thought Microsoft was on a downward slope—giving Windows away for free, or nearly for free (I personally download it as an ISO and use it in a VM at no cost, and everything works fine). But their cloud and AI services have breathed new life into the company. Microsoft is the largest shareholder in OpenAI, providing the cloud infrastructure needed to run their models. They also grant access to their search engine, allowing the models to retrieve data from the internet. Since OpenAI is not publicly traded, it remains insulated from stock market volatility. If there’s no progress in a company, there’s a decline. Year after year, employees demand more and more for their work. Inflation is driving up the prices of everything. If the IT sector fails to create new products, it’s game over for them. Blame capitalism.

Wow. What a load of nonsense.. This guy is a lamer.. To disable vulnerability in PHP without (temporarily) patching anything you just need to modify php.ini: ; CVE-2026-109619 session.upload_progress.enabled = 0 ; CVE-2026-109619 – PHP 8.5 only fatal_error_backtraces = 0 ; CVE-2026-109618 memory_limit = -1 ; POST size limit post_max_size = 8M upload_max_filesize = 4M ; Variables limit max_input_vars = 1000In some other config files: php_admin_flag[session.upload_progress.enabled] = off php_admin_flag[fatal_error_backtraces] = off And Apache: LimitRequestBody 1048576 or Nginx: client_max_body_size 1m; This removes known attack vectors for PHP. Which was completely unnecessary, as it would have sufficed to prevent out-of-memory errors by simply adjusting a few configuration files (as in the procedure above). Cloudflare did exactly that, and their proxy servers were immediately secured. To reach the servers behind the proxy, hackers would have needed to possess their actual IPv6 addresses, and the servers would have had to be misconfigured (as they still are!). They saw it because they had implemented limits on GET and POST data. They blocked this data to prevent it from reaching their clients' servers, ensuring it wouldn't trigger out-of-memory conditions or cause a longjmp() to be executed. From your later post about "tech bros", I conclude that you didn't understand any of it, since you blame these "tech bros" for something.. And they protected your ass..

It’s easier than you might think. You give an LLM/AI a piece of code—even code that it generated itself—and ask, "What problems do you see here?" It then starts looking for them. It’s the very same code it provided earlier (after you asked it to write something), yet it’s full of errors. The AI might point one out, saying, "Oh, I see an error here, because of such-and-such..." You reply, "But you generated this code; I didn't write it." The AI then responds, "Oh, sorry, you're right—here is the final version of the code [...]"—to which you point out (if you know your stuff!) that there’s another error. This whole process can go on for hours.. Running automated code analysis quickly uncovers errors in the code—especially when the source code is available, as is the case here. For humans to perform this same task, it would require thousands of people per project and entail enormous costs. Every electrical or mechanical device has some issue that manifests only under extreme conditions—conditions in which most people never actually use it—causing it to stop working. The same applies to code. Such bugs only appear in extreme scenarios, because the obvious bugs were already caught at an earlier stage. "PHP uses longjmp() for out-of-memory (OOM) errors to exit the current context and abort handling of the request. Some contexts may not properly account for this scenario and leave request memory in an uninitialized state before performing an allocation. The subsequent request cleanup procedure may read or branch on this memory. Multiple OOM strategies can consistently trigger this condition and lead to remote code execution (RCE)." "PHP has various global state that temporarily points into request memory. In some scenarios, global state may not be correctly reset, or may be repopulated after being reset. The global state may then refer to request memory even after it has been freed. Subsequent requests may read or branch on this stale state before it is fully initialized. Multiple strategies can trigger this condition and produce remote code execution (RCE)." How is it typical scenario? Your car cannot fly or swim either.. direct your complaints to the manufacturer.. Legislation has no bearing here. In a project developed for free, there’s usually just one programmer doing everything—and they don’t necessarily have to be an expert in every aspect—often working on it during their spare time, outside of their regular job. If someone doesn’t want to use it, they don’t have to; there’s no compulsion. The catch is that those who do use it often feel they have no choice, as there’s no easy alternative. In commercial IT, software is written in a way that makes it easy to replace the programmer; they spend more time on paperwork and documenting every function they write. Employee turnover in IT is very high—it’s rare for someone to stay at the same company for more than a few years. Once they leave, no one touches their code because, in reality, no one knows what’s actually going on inside it—unless it was thoroughly described in documentation or code comments (which, historically, was rarely done). Personally, for instance, I almost never write comments or waste time on such nonsense. People get bored; they want a higher salary or a new challenge rather than doing the same thing over and over, so they change jobs. According to Google Gemini, the average tenure at a single company is 24–36 months. When someone like that comes along, no one knows what was in their code or how to fix it. Often, the code ends up being rewritten from scratch, because analyzing it to fully understand everything would simply take too much time. Code generated by an LLM/AI—now that is truly buggy! You could find fault with every single line... Back in the days of floppy disks and CDs, much greater care was taken to ensure code was free of errors—because once a product was released, it couldn't be updated or patched via the Internet. Once a CVE is in the vulnerability database (21 October), anyone who can read English can recreate an exploit; there’s no need to analyze the code of the patch—though "patch" isn't really the right word here, as it involves shipping an entirely new, fresh php.exe. and not something like Windows-style patches. In the case of Windows, analyzing such a patch—had it not been included in a CVE—would actually make sense.

Temperatures in Moscow are usually lower than in Ukraine. I have no doubt that this winter, the Russians will be heating their homes with wood and coal.. provided that the trucks carrying them can travel faster than the fuel runs out.. This will be the worst winter for Moscow since 1945.. but they brought this upon themselves.. ps. Funny, I just finished watching a 2017 film called "The Death of Stalin".. They shoot Beria at the end.. https://www.imdb.com/title/tt4686844/

In its early days, progress consisted primarily of increasing processor frequency. Doubling the frequency meant doubling the speed, and so on. Until late '80 ALUs and FPUs in either Intel and Motorola CPUs were separate. https://en.wikipedia.org/wiki/Arithmetic_logic_unit https://en.wikipedia.org/wiki/Floating-point_unit So if one wanted to have floating point math, he or she had to emulate it with integer ALUs, which is a very slow process. After '84 in the case of Motorola, they could invest in 68881 FPU coprocessor. The first Motorola with built-in FPU was 68040 released in 1990 https://en.wikipedia.org/wiki/Motorola_68040 The first Intel with built-in FPU was 486 DX released in 1989 https://en.wikipedia.org/wiki/I486 (In 486 SX it was disabled) For over two decades, the main driving force has been increasing the number of cores. If you double the number of cores, you get double the speed, etc. But this requires that your algorithm scales well in a multi-thread environment. Which isn't always possible. Sometimes it's never possible, because, for example, the result of the previous calculation is a parameter to the next function. If the goal is to calculate something as quickly as possible, instead of writing a software program for a standard CPU, you can implement the algorithm directly into the hardware logic using an FPGA. You're essentially creating your own private processor instead of making software. https://en.wikipedia.org/wiki/Field-programmable_gate_array The cheapest models on AliExpress cost $30-40. Search for the Tang Nano 4K/9K for example. Is it even a computer if you have, for example, a cluster of machines doing one task? After all, you can ask people (billions of them) to download some app on their phones/PCs and send them a task to do when those phones/PCs are in idle mode, and can you call this a "supercomputer" - a mesh, grid, of machines connected via the Internet? Hackers, for example, mine Bitcoin. Someone visits their website, the browser starts calculating their code, multiplied by thousands or millions of visitors.. My processor in this laptop has 0.5 TFLOPS. I'll turn on the second laptop and we'll have 1 TFLOPS.. ;)

ChatGPT v6 - what a "beautiful" total disaster... It is completely unusable.. One browser core is constantly occupied at 100% and the more you write, the faster you will reach a wall after which LLM will not respond at all and will block completely.. Do you experience the same issue on your PC, Chrome, or other browsers? I'll have to check if Chrome and/or Chromium will also behave the same way as my Firefox.. Besides, in the previous version of the UI we had nice fields with code in e.g. C/C++ with copy buttons etc., clearly separating the area from the rest of the text, and now the code has been merged with the response, copying has disappeared and everything works hopelessly.

Copy your entire message to ChatGPT and it will tell you were you made mistakes..

That's pretty normal. The thing which makes a difference if they are "smart" i.e. they have their own OS, and have Internet connection.. If you can read websites on them, they must have to have Internet access.

Do you use a web browser from Microsoft? I doubt it. Anyone who installs a new system immediately downloads Chrome, Firefox, or Opera. To "know you," they need you to use their web browser. Google knows you because you use their Chrome.

You’re too fixated on that novel and fail to see that the difference between a PC, a smart phone, and a smart TV is merely an illusion. Every smartphone is essentially the "TV from 1984" that you carry everywhere with you, making it far more powerful than the one in the book (which just sat on a table). Information about where you go and the routes you take (along with that of everyone else with a GSM+ phone) is widely available in mobile operators' databases (and to anyone with whom they share that data). Fragmentary data may also be available to the developers of apps you have installed (since this data is derived from the Wi-Fi hotspots you connected to). A Smart TV runs either Android or some other OS that allows you to install apps - so you can also install something hack-related. Turn on the audio recording on your phone, place it somewhere - like on the balcony - then walk away while counting aloud; later, you can listen to the recording to see just how far you got before the sound faded out. When I ran this test, there wasn't a single spot in the apartment that was safe.

Every version of Windows starting with version 10 has a GDID. https://es.wikipedia.org/wiki/Global_Device_Identifier When you start up a Windows computer and connect to the Internet for the first time in your life, you are assigned an identifier that stays with you forever (unless you format the drive). In reality, there are at least two GDID identifiers: one assigned to the machine and another to the user. Consequently, when another user is created, Microsoft knows it involves the same machine running the same operating system. Every time someone connects to the Internet, this data is transmitted to Microsoft servers. Your IPv4 address - or, even worse, your IPv6 address, which is Universe unique - is revealed by the TCP packet data. If you turn on a VPN or Tor, Microsoft knows that you previously had IP X and that it later changed to IP Y. So, it is clear that it is still you. If you bought a computer or laptop with a pre-installed operating system - which is practically the standard - and you’re the sort of typical "useful idiot" who’s actually happy that you can just click and everything runs automatically, then Microsoft ends up knowing exactly which store you bought it from, the store's precise address, the time of purchase, and so on. As a result, they could get hold of your credit card details and your real name. If the store had surveillance cameras, they might already have a real photo of you. If, at any point, data from your Wi-Fi or Ethernet card is sent to Microsoft servers, even reinstalling the system from scratch won't help you cover your tracks. With every Android phone, unique data - such as the MAC address and other information - is definitely sent to Google. In your Google Account, there is a panel where you can find a list of authorized phones and other devices (those used to log in to the account); you can even view the MAC addresses of your past and current devices there. Google knows your phone has been stolen before you do, because someone will try to set up a new account using the same MAC address as yours (which is in their database). And if he tries to log in using his existing account, they might be able to find out who stole your phone. Every installed application (whether on a computer or a phone) can have its own unique identifier. They don't necessarily have to use that GDID; it could be something completely random. If you factor in automatic app updates, the developers know when you turn on your computer, when you go to sleep, whether you're using a VPN or Tor, if you're flying on a plane with Wi-Fi, and how often - and where - you travel. If you used hotel Wi-Fi, they know which hotel you stayed in - i.e. what standard you can afford. Any application (including the operating system) that performs auto-updates can be used to install spyware on your system at any time. Every version of Windows and Linux remembers the Wi-Fi networks it has connected to in the past. It stores not just the name (SSID) but also the MAC address (BSSID) (global unique). This information is buried so deep and in so many different places that the best solution is to format the drive. Every Windows and Linux system stores the globally unique identifiers of USB devices connected in the past. For instance, if a friend drops by with a flash drive, copies some data, and leaves, and it later turns out that the drive had been used by a hacker, that data will still remain on your system. The only sensible solution is frequent disk formatting or, if using Linux, running a live version such as Kali Live.

If you generate private and public keys and then use the RSA algorithm without padding - that is: openssl genrsa -out private.pem 2048 openssl rsa -in private.pem -pubout -out public.pem openssl pkeyutl -encrypt \ -pubin -inkey public.pem \ -in test.txt \ -out encrypted.bin \ -pkeyopt rsa_padding_mode:noneEncrypting the same text with the same public key yields the same output file; this means that very short text messages in known languages can be recovered using a brute-force method based on a dictionary of words or sentences. A key aspect is adding a random padding: openssl pkeyutl -encrypt \ -pubin -inkey public.pem \ -in test.txt \ -out encrypted.bin \ -pkeyopt rsa_padding_mode:oaep

@Trurl Me thinks, you never did it in your life.. So, let's do it. On Windows download VirtualBox, download some Linux ISO file, for example Ubuntu or Kali Live e.g.: https://www.virtualbox.org/wiki/Downloads (170 MB for Windows, 110 MB for Linux - but if you have Linux you don't need to install VirtualBox in the first place) https://old.kali.org/kali-images/kali-2024.4/kali-linux-2024.4-live-amd64.iso Create VM with 4GB RAM, disk space is irrelevant - you will use Live Linux i.e. without installation (it will be a way quicker). 10-15 seconds to start, instead of 15-30 minutes to install it. You need to have openssl installed. Kali and probably most of other distributions have it built-in so nothing needs to be installed. To create private key with RSA algorithm you open Terminal and use: openssl genrsa -out private.key 40964096 is the number of bits / 8 = 512 bytes. Example private key file: -----BEGIN PRIVATE KEY----- MIIJQgIBADANBgkqhkiG9w0BAQEFAASCCSwwggkoAgEAAoICAQCiC9MUyjygZWGB F9g9EysaNp46O0NEOUTJmlnzV6WlXvKqn1ggPGmFzGxUB4UPTOrFdMGGrZRPwPLP 6X7VZTRZTrEiu48qrkP/dCOX79XHCcgK0iBTuDPjM9KZLiovg3EjS1CrWC4swcKx UF0wpL7QwQIlfKyYThL2FPmVIf4hPWDvvaEYb64C+onxTlqVPuHl4K3va//tXl92 tdVR9mzrVLAsneyj5aUlQUiVgl0eTzaRUG4TYU5bC9SW406A6bA2ev04ChrhG5oI wa62Gf/C7o6ZYnlmrjTw1yMfzrDwBsuyr+RIvH6YubM8ISArwhPQ4aQMR/iq7SWA PYnQi3JjkQwbRW0QmI3wgFMmW3ML/kmvd6Gqn9x49nuo3yarttUPVkLhtF0pDNLY 1UawXpJ2WK4X1hL1zJ0HLCQBAzY4jvZG3TEI6cL7RueVCE7e8TwwANPVVB9J/YEj PX6J2sf9uvIqktKaJunZDxR2sliTGvZyJj9+8xIss1U0lYoEO5K/XiYua3g7PBLC Se1o4w03dUgb/kwSgnSWahv1Z7zhqea+e2WOTH4YWY0Cy+NBFYQ2WmBcVHNKyij/ 5MSVisbSAevZSFY+5UEjsOiFz4nyapXpg04q7p5WwmxVlRuQ6vlDn7DeOSiRS7HY uyZZuCJOqO2jdO2vWX1wWdjj27MBNwIDAQABAoICABulEccns00XzpkT5M+SkBVi QP8lMhYXAQ21t+fESZ2bQ6HpGkMxmzz4wXhaahBQqnKyNvvPbBBpFeNzbbBhGhus SZrnzYZWAYK4oQPgJiPVHtv9ugceUzfNFP4N+qNtdKU1SZdZrVL0YU8XqzZLvMPn UYLssJMxKpnxvRx8lpq8Gmf7/VoqX1Dhxhu1Js30UtyBRpskXR51LasWXUSk2jMR mRabgI37dP+UkhnzighMMj57LrQ+svvkWkm9kJx8gZPhyoY/bxN9PnuE81XcvJ3Y TdqMCrMCQETgJv1bqSLhvemHiVYBVkWe1J2K9y6P5wqJHA/+TUV3vKCSnwrFZ2eg GLw+OhGO+v321qv1wRXJ4qAoF4812P1XZRf7WV3A3S1MX37lGqxu8PH9VFhbjmX5 9tTehwdXtqjNHZRShjh2yz/AHJVdF1oRrwIf7GyyN4aE+/aRU0r6MhuTbJan1tWW RbC8sC0DYQI3/BQlHfoD+yRHmeMwW8uZgVhGfQtT5WYOvPA8dnn4qd65GrJAtL9w Y6a0PWDvBV75+eJ9Ho+k8BfINHP7T8fejJpGUjeEiJrbv3qB+wMmF0tjFqymHgXr /8ezN1Eb14g9N1ZTzryqEdp6MEz64SqRI0wZFmsc1mTcPiuPVHqYJrvEmu37NcmI dKSC69tDrcaxcw1HivwBAoIBAQDkcmU04pucCZUzIx1uoJ1IG1jBjaVvcaxiDlu/ pamO8x+GO1bNRHQfCrkwpl92NZ1RrOLoDK3haPiNArK7hudmCihEaDbDNYhxbD/l eoYYaUSjRqE/gu243s4VB2pWG76bmjGp5QwPZl9i1zPoJmRMKb1Uzfp7moK9O2RP UGWMFgfCllv6AEY935GL4BXFIy/pxCA069GozQIIxs5W4BQcfZK+AqVN7sG2ATyg gIgu8T6avMi/SvbrLUdv6oe76oOegcHee596Z1sU4Z268opBFEfb1pzxCszNKkDB H5p7+yyczc/0sdMhQwhlC9J1xLD8ci68cVJcqzbETkQzQXU3AoIBAQC1lzgWDUiw 5gN+0FR5t7vLi78Ai7gKL6ECagqQEbE4lmhC3NhQsilhYUtzQKvD56AzPBQq2k9p eAbLxULf+7mATcwRV0xQRwOCAkq7HclWM+25PjgRxlXMygqSAVyqlyvrvSdqnlxz K3fo7rWxqLqkDB0V5tLZVGFtnPA1QEaRVoyJBtpMloyFrHp7mGeFM9uIpsAhhw9N qrGTz/BI4P+d4Tnt2pyGFM/5V3s6tDR4WFeq4vYEj/2pU1BEK9wI45wFJ32u3X0l /BfLCGTU2yvVO8h2sKT8h/ole3qP5kwnc2ASjyplev3xQGRYpvE5MHUbLH8RkdMu XpBKOkXXoNQBAoIBAQCt9ttRSZu8WKPQvFXmb9M2HPUa9/XahxMB3PNLXylfHY+N FdUFHbLcDLt368pYZ320u9ZTZYxoUhqsrzhSi+ndD7EYXtANGdAcp5WE+CfzoP5O 5DNY/ShwluMSqoH0rs4ZCnP1b+IePBQ5Xa4+hDXutf5i/TRZOhGDbx31NEEQS5p/ OSczm0/iQEVC4lNRu3j6KItqIYaYZ/27l0F8cMl/5LSBk9w0G7RSGGEFzLtoG/hT t52b5sdf0ixLcqsIlA7E/wUWmseYW+rFXQZTXe2gn4QVSjpbaKELP7juGFX5Qg34 HtbU/THNa5GR+U28txscMwuaHyXcQdnD5eITsi5DAoIBABFahTKeTF8TbCwXTqtf IqQ+E5/CunDGwnVPmvjPqAcHYKlbH5T+obO7r5ExEfkjzox25lp+CG7E2KlqjYyI Kt8XEDRsCk/1qWEoROoC27XQOpykVIlmUdbDTCVN2gQ3dCwQJI8s2RrsgYYSIez7 efKpcbIhH+9DABiQkegU32mOQtnpfL0q+fzgUcgz9VimbwltHhlWJeE/Ob8tdQIR 9nupBXc0mBa8MG/cq1sd05HYp3AeRnKDLDOdUaNiKaHz1nornvgyFH6ny8JMaA5r b5HuJ6coFGxJtl8UoM7VrFtszDaZ1exix5fhlnVeZtkXR2m0Y+RS8kKRKOrfDMOh jAECggEAX5f1KbYaZkx72M3MAfKZh237bgKPuJ+a7CtgU62ZknHqnFsoNvHD1uTz T/fxWW9xgyBX3tcaOse+PaYQliIiSu4A0HnN0sawKeFvLOIet3ushtenTgkCBbGl c8jWcrJe3Yk0bf61Vekse+kW6GAAQffJU0/85GWbHyU71xHmpZfgm8G4rdt7iRry ygjargCzUFzbIkKaq6ctXjlmxwp3PFYsn+65Yo4EYQ5Fsp+YYc8BvdmB/h0FC34x +LwWzWsqBDCDqvgFS1KX1Zr3aZdCn5NejDzESj0dEf4IWqc9JLkybDbwJ26o0rgv TNlh3K5wYBLgldBnXcy7gjTJoKP1Jw== -----END PRIVATE KEY----- Public key is derivative of private key: openssl rsa -in private.key -pubout -out public.keyi.e. private key is input parameter to public key generation procedure. Example public key file (matching the above private key): -----BEGIN PUBLIC KEY----- MIICIjANBgkqhkiG9w0BAQEFAAOCAg8AMIICCgKCAgEAogvTFMo8oGVhgRfYPRMr GjaeOjtDRDlEyZpZ81elpV7yqp9YIDxphcxsVAeFD0zqxXTBhq2UT8Dyz+l+1WU0 WU6xIruPKq5D/3Qjl+/VxwnICtIgU7gz4zPSmS4qL4NxI0tQq1guLMHCsVBdMKS+ 0MECJXysmE4S9hT5lSH+IT1g772hGG+uAvqJ8U5alT7h5eCt72v/7V5fdrXVUfZs 61SwLJ3so+WlJUFIlYJdHk82kVBuE2FOWwvUluNOgOmwNnr9OAoa4RuaCMGuthn/ wu6OmWJ5Zq408NcjH86w8AbLsq/kSLx+mLmzPCEgK8IT0OGkDEf4qu0lgD2J0Ity Y5EMG0VtEJiN8IBTJltzC/5Jr3ehqp/cePZ7qN8mq7bVD1ZC4bRdKQzS2NVGsF6S dliuF9YS9cydBywkAQM2OI72Rt0xCOnC+0bnlQhO3vE8MADT1VQfSf2BIz1+idrH /bryKpLSmibp2Q8UdrJYkxr2ciY/fvMSLLNVNJWKBDuSv14mLmt4OzwSwkntaOMN N3VIG/5MEoJ0lmob9We84anmvntljkx+GFmNAsvjQRWENlpgXFRzSsoo/+TElYrG 0gHr2UhWPuVBI7Dohc+J8mqV6YNOKu6eVsJsVZUbkOr5Q5+w3jkokUux2LsmWbgi Tqjto3Ttr1l9cFnY49uzATcCAwEAAQ== -----END PUBLIC KEY-----To encrypt a file, so it is not possible to read it without private key, you use: openssl pkeyutl \ -encrypt \ -pubin \ -inkey public.key \ -in file.txt \ -out file.enc \ -pkeyopt rsa_padding_mode:oaep \ -pkeyopt rsa_oaep_md:sha256 To decode you need your private key: openssl pkeyutl \ -decrypt \ -inkey private.key \ -in file.enc \ -out file.txt \ -pkeyopt rsa_padding_mode:oaep \ -pkeyopt rsa_oaep_md:sha256 What knowledge do you have after a real life testings? Public key must have enough information to be able one-way encode a file. Private key has enough information to either encode and decode a file. Public key is derivative of private key. ps. Generated example private and public keys are base64 encoded to be readable by a human. Ask ChatGPT/Gemini how to generate binary files instead, if you want to use binary and eventually write your own RSA encryption/decryption.

When I want to transfer files between phone and PC, I set up web server on the phone or PC. On the phone you do it by installing Termux. e.g. https://f-droid.org/packages/com.termux/ On the PC (easier) you do it by installing Apache or XAMPP (even easier): https://www.apachefriends.org/pl/index.html or if you have Linux, I simply write "service apache2 start" (web server folder is in /var/www/html ) Then you make sure you have both devices in the same local network e.g. make Hotspot on the phone and log in this WiFi from PC, or vice versa. To learn what is IP address you use ipconfig on Windows, or ifconfig or ip addr on the Termux. Then you put files in correct folder in Apache. Ask ChatGPT "what is apache folder on Termux" or "what is apache folder on Windows" And then you visit http://ip-address (which you got from ipconfig/ifconfig above. If you copy'n'paste this reply into ChatGPT/Gemini, they will explain you my procedure with more details. Of course this is procedure for remote transfer.. The easiest non-remote way is by simply plugging USB cable.. "Syncing" means "sending all your files to the cloud, and then redownloading it from the cloud".. i.e. if you have poor Internet, rate limited, you can't do that.. It is plain stupid.

Either Google and ChatGPT gave me "Somerton Tunnel". Which answer is correct? IMHO, you shouldn't use "NE", but write full "North East". "NE" is too ambiguous. We don't know if LLM understood "NE" correctly. The tokenizer needs to convert human-readable words into tokens. "NE" is not an English word. "NE" may be a word in some other language, with a completely different meaning than you think. There is no weights database ("~100B parameters") per country. There is just one weights/parameters database for all languages. We don't know how in "Google AI" it is implemented internally. ChatGPT tokenizer is one for all languages. It does not recognize languages. What Google gives if you skip this "NE" and asks for "What is the name of the mainline railway tunnel in Taunton ?" BTW, if I cut it to "tunnel NE of Taunton" I got this: You realize that such questions are not possible to be answered by pre-trained LLM from internal weights/parameters database, right? It can't fit all in "weights" aka "parameters", right? "AI overview" uses Google search engine, and "reads" these articles which you have straight below your "AI Overview" answer. So, for example, if somebody is buying ads for "xyz" because e.g. earns money on selling tickets for historical place, such articles will be on the top and in many different forms (social media, and online media, for example). "AI Overview" will read them, and create something from them. “AI Overview” is not the right tool for what you want to use it for. Try: Gemini ("Google AI" in chat mode). ChatGPT They work on different principles than "AI Overview"..