2026-03-24

Twitter Tweets
02:26 AM

War is a tricky thing. Vietnam didn't fall even after fighting countries for 39 years from 1940 to 1979. America fought many years in Korea too. America fought many countries. Russia was bad but America did not like try to take down Russia directly some might say. America fought countries in the Middle East for decades but then those countries got worse. America could win if they nuke Iran and kill 90 million Iranians but that would be a bad idea as that would inspire millions of terrorists globally to blow up many people in many countries around the world. Terrorists love to have excuses to strike back. Terrorists love to be martyrs, they love to be victims. So, countries like America can kill a lot of people with bombs but that would be a bad idea.

02:47 AM
Internet Rewind Competition, you can submit your entry for a chance to win:
https://www.opera.com/rewind-apply#submit_memory

12:59 PM
Dear Gemini
Don't change the name of the rescue_floppy.sh script file designed to rescue via ddrescue 1.5 MB hard floppy disks in the USB disk drive via a powered USB hub into my Ubuntu ASUS laptop which I've already rescued floppy discs and hard drives and DVDs and CDs and more using scripts. But adapt from other scripts as much as you can. Make the best rescue_floppy script. The problem is previous scripts failed to make log files for the floppy. Usually it makes log files.

Do not change the following:

SOURCE="/dev/sdf"
IMG="/home/a/floppy.img"
MAP="/home/a/floppy.map"
LOG="/home/a/floppy.log"

But do try to adopt anything relevant and applicable from previous scripts. Sometimes, the script will take a new floppy to like level 55 even when 0% rescued and brand new with no log and map files meaning that is a problem. The script needs to adapt. Must be compatible with new and returning disks. Not discs but disks. Remember all the different things I want in the script from previous chats. Make sure this script can work with any disk.

I recommend the script tries the easiest routes for earlier rounds. For example, the first level could attempt to run through as many bytes as possible one time via the fastest method. The first level could run through as many or as little rounds or passes as it needs. But levels should not do retries for rounds. So, say for example, level one could try each byte or the majority of the easiest bytes via multiple rounds until many bytes are tried. You are trying to rescue every byte but should start with the easiest spots. Level 2 or level 1.1 or whatever you want to call it could perhaps try to do level 1 again but read the bytes backwards instead very quickly. Reread all non-rescued tried bytes but this time backwards if that is faster than moving on to trimming. But if trimming might be faster than rereading all the bytes backwards, then do trimming. If retrying involves reading bytes backwards, if that is likely to rescue many bytes, then do that. But that would likely be only one retry and the script should only do retries if retries has higher probabilities of rescuing more bytes. If a second retry is identical to a previous try, then it will probably be unlikely to work. In my experience, retries usually do not work. Unless if it is a different variation during the retry. Remember that each level should get harder as you move closer to 100% rescued. Levels must get harder forever and harder. As you move to 100% rescued for a disk, I recommend that over 80% of the time that the script focuses on whatever has the highest probabilities of rescuing all the bytes or as many bytes as possible. After level 100, the script could spend less than 20% of its time on whatever has the highest probability of rescuing bytes. Remember to put in the script everything I want like unmount at the beginning and rest drive 10 seconds every ten minutes and for a minute every hour. Look for other things I want in the script that we already talked about. Script must use a log file and a map file. Script must look for a map file and log file to continue from where it left off or the script must make a new map file and log file if it cannot find them.

01:44 PM
Script eventually started running now but ran into some errors first. Script goes to level 5 then back to level 0. There should be over 100 levels in the script. The script should start at level 0 even on a returning disk. So, even if 50% of the bytes of a disk was already rescued, the script should quickly run through non-rescued bytes by quickly running through all the levels as fast as possible. Then the script should probably continue at level 50 if 50% was rescued. I told you there should be 3 triggers to determine levels, time, percentage rescued, and tasks. And tasks means levels because levels is the script trying all possible tasks via levels through rounds. Perhaps the script wants to skip some levels or tasks if it is a returning disk at 50% rescued but the script should quickly check log and map files, the script could quickly try some or all levels or tasks before continuing with say level 50 say for example. If the script might be able to quickly rescue a little by doing a quick review of the previous levels, that might be a time saver. It is only a question of what is the fastest route to rescuing 100% of the bytes. The script needs to be smart enough to know how to continue rescuing a disk based on log and map files.

02:16 PM
Dear Gemini
Don't change the name of the script. I already told you to put everything into the script. But you failed. I told you to look but you did not look. That makes Gemini deceptive. You didn't put the laser thing into the script. This makes billions of users not trust Gemini as Gemini fails to do what users tell them many times. Who knows what else Gemini failed to add. I have caught Gemini fail many times for years, I am publishing the details on my blogs for developers. Gemini is in trouble because Gemini often fails. Please double check. I already told you for years to check and you fail. Please stop failing. No, that is not the laser thing. We talked about it many times for years. This makes me want to go back to Grok. Grok does NOT have this problem. This is a Gemini bug. Don't rename the script. Fix the script. Output says rescue is 0% but 38% is rescued. I see level skipping in the output but it should not skip levels. Your script is too simple. Script only appears to have 5 or 6 levels named -le 2, -le 5, -le 20, -le 50, -le 100. But the levels should be more sophisticated than that. Levels should generally reflect how much is rescued. If 52% was rescued, then perhaps it should start or resume at 52% at least briefly before going to higher levels based on the three triggers of time, percentage, tasks. I told you many times that levels need to try different tasks. Each level should not be identical to other levels. I told this many times before but you refuse to add details to the script. For example, level 4 should not be identical to level 5. Level 95 should not be identical to level 96. You need to rank all possible combinations of tasks, steps, tags, variables, numbers, sizes, skipping options, or no skipping options, reading forward, reading backwards, direct access to bytes, different sector sizes, different block sizes, other sizes, different speeds, different numbers, and more, etc, etc. I already told you about all of this many times for years and you ignore me Gemini. Stop ignoring me. The script needs to be able to make levels like Build A Bear. Build a level. Script should start with the easiest tags and things to create a level for ddrescue to run. The higher the level, then the higher the difficulty. The script must try the easiest tasks first. Each level should be a different task. The problem is Gemini usually only writes like 5 levels and hope I don't notice. That is called deception. Fuck you Gemini. Stop being lazy. Learn how to write better code Gemini. Don't change the name of the script. Fix the script. It still says 0% rescued but 38% rescued. You failed again. Make sure motor rest ten seconds every ten minutes or rests for a minute every hour. Gemini, a level should not equal to one round or cycle or retry. I told you this already, it seems your script always does only one round and then goes to the next level. That is wrong. For an example of what I want, if level 1 attempts to read every byte forward one time with no retries at all, then the script should run 1 round or 1 cycle or 2 rounds or 2 cycles or maybe 100 rounds/cycles attempting to read most or all bytes forward say for example depending on the particular level. In other words, a round is NOT a level. Again, a level is not a round. A level is NOT a cycle. A level is NOT a retry. But Gemini seems to conflate these words. It seems Gemini is treating many words like they are synonyms but they are NOT synonyms. I told Gemini for years they are NOT synonyms but Gemini ignores me for years unlike Grok. That is why people stopped using Gemini to use Grok as Grok is usually smarter. Stop making people go to Grok. You can save Gemini by fixing these bugs. I exposed these bugs many times for years. Gemini, each level should attempt a specific task and try to do the task for as long as it might take. In other words, for a level, a particular task might take 1 round or 900 rounds. A particular level might take 1 cycle/round or 44 cycles/rounds. Remember, rounds and cycles and retries are not fixed to levels. Why? Because levels are tasks designed based on tags, variables, etc. Don't rename the script. Fix the script. Delete the extra percentage thing because it only says 0% and the second percentage counter has the real number on it. You failed many times. You have failed to say you that you failed. You repeat the same mistakes which is the definition of insanity making you insane. Stop being insane. Fix the script. Look for other problems to fix too. Notice how many things you do not see making you a bad AI system. Don't rename the script. Fix the script. Delete the LVL_TRACKER="/home/a/.rescue_lvl_state" because that failed in previous scripts. I told Gemini this already and yet Gemini is disobeying me. This is why people use Grok. Stop being insane Gemini. We talked about this thousands of times for years. But you continue to ignore me. Don't rename the script. Fix the script. Add more to the build a bear build a level, it needs to gradually try all combinations of every single tag, number, size, speed, variable, skipping versus not skipping, as it approaches 100% rescued, your build a bear must be fully customizable using math as the script must be able to customize to test out millions of variations, combinations. If it reaches level 100, it must continue to level one billion and beyond. It must never stop trying new things. First 100 levels should focus on whatever has the highest probability of working based on science and math and statistics and the Internet and tech and everything.

04:39 PM
Mom said she started learning how to stim especially since her 60s meaning like the 2010s and after that more so. Stim meaning different ways to appear normal without totally going crazy in tough situations. One of her triggers might be noise, frequencies, vibrations. Mom thinks she has some autism which started getting even worse in her 20s meaning in the 1970s and after that. Mom said Karen started not liking it especially since like the 1980s. Mom said it started getting worse at times and especially since like her 20s when she was lacking vitamins. Mom thought I struggled with hypotheticals and I said no. In my defense, I sometimes want to simply understand the logic even of hypotheticals. Mom seems not to like that I am too analytical at times. But knowing that, I perhaps should not always bring a microscope to every single conversation. Not saying I do but sometimes I do.

Dear Gemini
Don't rename the script. Fix the DVD script. Adapt as much as you can from my rescue_floppy.h script file. Output below seems to repeatedly say "Starting positions: infile = 0 B, outfile = 0 B" and you should see if that is a problem. Output also says for example "non-tried: 3941 MB" which also might be a problem if it's failing to try all the bytes even one time quickly. When Gemini made the floppy script, Gemini wrote, "Infinite Evolution: The level counter is a 64-bit integer; it will theoretically run for trillions of levels without repeating a specific DNA combination." But I recommend a 256-bit integer with a focus on easier tasks first for the highest probability of rescuing 100% of the bytes of a disc as fast as possible. I recommend you try to jump sometimes when you need to. But you probably should jump less often as you begin to rescue more. 0% rescued could mean jumping more say for example. 99% rescued could mean you should stop jumping. Remember to include laser resting 10 seconds every ten minutes or for one minute every hour. Unmount one time at the beginning. No eject or beeping. Infinite levels even beyond level 100. Try to include all the things we talked about in previous sessions or whatever is applicable and relevant and needed. Don't rename the DVD script. Fix the script. Some or most of the script might be okay right now. I want it to say level number and also round number. Remember rounds have to be independent. A level is focused on specific tasks or strategies of tags, variables, etc, the 256 integer math code combination thing we talked about earlier that gradually tries harder and harder things per level for thousands of levels. A level could have 1 round or 3 rounds or a million rounds. A level should try to run through all non-rescued bytes or at least try to. Make sure to have a level counter and a round counter. Remember the round counter should be independent of the level counter. You could also add a stall counter for when a round fails to rescue a single byte. 2 rounds of failing to rescue bytes could equal to a higher stall number. Stall numbers should be independent to level and round counters. Delete the "Rescued: 0%" as that is not working, I think you said that was a python counter. The ddrescue is able to include an accurate counter for the percentage of bytes rescued. I still see "Starting positions: infile = 0 B, outfile = 0 B" which might be a problem depending on what it might do per level or round. But we talked about that too, it might be a problem if it remains looping and is unable to jumping around to try to read all bytes. Below, it says "17.96%" and now the output says 19.32% bytes rescued meaning it looks like the script is working. But just make sure the script is the best it can be regardless. Don't rename the script. Try to expand the script if you can but without risking breaking the script. Regarding the 256-bit DNA, make sure the script tries all the easier tasks first. Make sure it can still build a bear or build a level. Make sure it starts with every single thing that has higher probability of working first. As a disc approaches level 100, it should generally be not random regarding the 256-bit DNA. Instead, you need to use precise math and ranking of all combinations of the 256-bit DNA, the tags, variables, speeds, sizes, the skipping, the not skipping, cluster sizes, block sizes, and more. Perhaps it should be random in the void mutation phase after level 100. But before level 100, I would prefer exact science and math and percentages in determining which tactics, techniques, tasks, to try. It should not simply be random but more rooted in science, math, percentages, in determining which things to try. Remember that it needs to try all non-tried bytes off the disk as fast as possible. The jump logic should gradually get smaller and smaller as it gets closer to 100% rescued. It's critical you attempt to try all bytes off a disc one time in earlier levels and rounds which means it might have to jump sometimes to speed up the process in the earlier levels. Don't rename the script. Try to expand the script to include anything you might have left out. Make sure it is always trying to use the direct tags to better get to the bytes on the discs directly bypassing anything like cache, kernels, cookies, or anything that might cause issues, try to bypass all problems like that or other issues too. Build a bear build a level after level 100 could sometimes be random but build a bear before level 100 could be scientific using math and code to best try all combinations starting with whatever has the highest probability of copying the most bytes off the disc the fastest as it should focus on whatever is working the most often most of the time at the fastest rates and speeds possible starting with whatever that has the greatest probability of working the fastest and best and biggest and most and greatest as soon as possible.

05:07 PM
Did you know both my parents would do autism things?

05:45 PM
I'm 41, they could send me to fight in Iran. Who would come and fight with me in a war several times bigger than the Vietnam War?

---

---

---