Skip Navigation

√𝛂𝛋𝛆

@ j4k3 @piefed.world

Posts
12
Comments
102
Joined
1 yr. ago

  • Most of the major developments that lead to the public stuff happened between 2017-2021. Transformers was the big one that made scaling a thing. Altman pushed in a stupid direction that caused a lot of the nonsense, like turning the name "Open AI" into an oxymoron.

    There are some aspects of alignment that point at political corruption and planning with nefarious intent that fits in with the present political bullshit too, but that is very complicated to explain in any depth. If you were to search the token vocabulary, you will find dubious elements are present in compound multi word tokens that disproportionately represent a single political camp, likewise with religious media, and science denialism. Much of that stuff dates from 2019 or before.

  • You need a model compiled for the architecture. I saw some for the RK35xx devices when shopping for hardware. I do not think there is software made to split up or run models in general on a NPU. The models must be configured for the physical hardware topology. The stuff that runs on most devices is very small, and these either need a ton of custom fine tuning or they are barely capable of simple tasks.

    On the other hand, segmentation models are small, and that makes layers, object identification, and background removal stuff work. Looking at your CPU speed, and available memory, it is unlikely to make much difference. You are also memory constrained for running models, though you could use deepspeed to load from a disk drive too.

  • Deleted

    Permanently Deleted

    Jump
  • I wanted to hotrod and play with cars. One silly magazine I read had an article about how automotive paint was a dark art few ventured into. So I started with that in my teens. I did about everything one can do with cars. I worked with heavy equipment for awhile just to get a bunch of welding experience, then went back to painting cars but doing bigger jobs and repairs. Eventually I did not like the state I was in so I started cycling to hotrod me. The cars started fighting back as more and more of me was lost without need to carry me around. One day commuting to work, two scrappy SUVs teamed up against me on a bike. I killed them both, but the broke my neck and back. I guess I wish I had not read that article... I have never really thought of work as a means to an end per say. I cannot imagine staying motivated by that. I am not the type to go along with bullshit or do what I'm told. Give me a responsibility and I'll do a far better job than anyone else and eventually I'll become the back office manager the owner trusts completely. I hate managing people. No task is beneath me. I do not care about narcissistic nonsense from myself or anyone else. I do what needs to be done and others are welcome to do the same around me. Those that do not are none of my concern. I am the employee I wished I could find when I ran my own business... or at least I was. Now, I am a shell of my former self, with 8 of my 9 cat lives spent.

  • On that level, maybe invert your mindset and look in Maker spaces. Search by hardware like ESP32. You will likely get better (different) results if you search for devices that target EE students instead of those that target Makers in general. Like it is well known that Texas Instruments will send free samples of most common chips requested, to anyone with a .edu email. Projects on hardware like a Beagle Bone tend to be more advanced than more common Maker hardware. While a BB is like half of a Rπ in terms of hardware architecture, if a purpose made device is created without all of the extra overhead fluff, it is pretty good. The STM32 H7 stuff tends to have advanced projects at the handheld gaming level. The Nordic BT BLE chips are usually more popular with the advanced crowd.

    You might look at the hardware commits for Micropython or Circuit Python for people adding DACs or other peripherals. These are likely to lead to their project spaces.

    I've seen someone doing a drive swap on an old iPod to SSD and a software chain, but I think that was still only doing the Apple compatibility thing.

    OpenWRT is not a bad place to look either. Any small embedded Linux device is likely to run on OpenWRT, so you may find something interesting just by shopping their hardware support and commit history.

  • I know, seems like a lot, but it really isn't if you just try it once. It is the same routine every time.

    FOSS has no marketing department.

  • When buying hardware in the present age, shop for FOSS software you want to run first. Then, clone the git in full. Finally, use the gource package to create a visual tree video that plays against the commit history. This will show you who is getting their pulls merged, how often, and how they contribute to the project.

    What you're looking for is who is consistent, and what they are using for hardware. It will always be obvious on larger projects. They will make little tweaks and changes a bunch between the hardware and software.

    You may see stuff like company employees and subcontracting devs come in and make large commits that support some specific hardware, but if you watch carefully, these are only a handful of commits, and then they never return. They likely had a checklist in a contract, completed it, and got paid. They will never return. Likewise, if one of the main devs gets a new device, they will shift to it and you're unlikely to see them make any further commits to the old stuff. The timespan between this transition infers much about the state of the old device support. Maybe just ask them why they switched and what is missing on the old stuff, or just cd to the hardware supporting directory and do $ grep -rin todo or similar types of stuff like code comments or words like hack or need.

    Hardware specs and advertising nonsense are worthless and irrelevant. Don't let highway robbers dictate your expectations. The only products that exist are those with FOSS support, so start with the FOSS and ignore everything else as criminal warlords. Who gives a fuck what products and deals the proprietary fascists churn out of Auschwitz or a Palestinian camp.

  • A disorder is a function that causes disruptive distress or deviation from nominal behavior.

    In abstract, I have posited a claim, and then shown how that claim is backed by associative social norms. I am attacking the normalization of anonymous negative behavior at a foundational level. I'm attacking the ethics of the developers that created this system in the first place. I have exemplified how this same behavior is in opposition to human social norms. I have shown its weaknesses in terms of political impact. I have posited a deeply unethical use case of why such a system would be implemented in the first place despite the malevolence. Finally, I have shown how it is destructive and harmful to everyone through statistical analysis using game theory.

    The abstraction is not targeted in any way at people with mental health disorders. I am showing how the feature itself is a disorder or catalyst for disorderly behaviors.

    I have actually tried really hard to remove any forms of bias or personal attacks from my dialog over the last decade or so. Like in this case, I'm actually arguing for positive constructive interactions in a more socially aware architecture. I want to remove the nominalized negativity. It was a mistake to make a space where people are able to abuse others, to manipulate, and to cause harm without social consequences as a feedback mechanism. It is a particularly sharp prejudice to experience when one is in near total social isolation from stuff like physical disability. Allowing people with no independent ethics to treat a space like this as a sadistic release valve for turgid eristics is simply wrong.

  • On pyfed, I see your overall "attitude" by default. This is the percentage of positive to negative voting interactions you have had recently. You are presently at 68%, which is rather low.

    The activitypub protocol is not at all private. Anyone with a server and admin account is able to see all of these details.

    Anonymous negativity is actually rather mental and should not exist in any democratic or ethical sense. You have a right to all information, a right to error, a right to skepticism, and a right to protest in nonviolent forms aka the right to offend others. Anonymous negativity is a violation of freedom of information and anti-egalitarian. Everyone has a right to confront their accuser with transparency.

    If you have something to say, you should have the decency of stating it. Downvoting is a mental disorder. It is like people that use four letter expletives to express themselves when they lack the intellectual depth to articulate their thoughts. It only really exists as a corrupt means of artificially influencing behaviors for commercial and political means.

    Is it ethical or reasonable to walk up to a stranger and give them negative feedback. Let's say you see a man exit his car to walk into a store. Should you have a right to leave an anonymous message on his car about the style of his shirt? Doing such nonsense will get you labeled a halfwit or worse. Take any real life circumstances and transpose this behavior. It is completely unethical nonsense.

    "Trust" as a mechanism, is the primary tool of authoritarians and fascists. That is trash. Democracy and community are built with open transparency and accountability. One is a coward. The other will engage the dialectic and has nothing to hide. My "attitude" is 100% now. I rarely downvote because the behavior fails at fundamental game theory and the prisoner's dilemma. Negative feedback is incapable of creating positive outcomes. It always brings everyone down. So if you are going to be negative, at least do so constructively in a useful way by articulating your thoughts in text.

  • Did you see I wrote that comment into a spoiler to shorten it in the post replies so it is not a wall of text? Sorry if my thoughts are hard to follow at times. I write in an abstract rough draft like format. I wrote several tips and how I approach a problem like this.

  • The floating thin sections on layers, lack the flow consistency to maintain temperature regulation of the heat block. The heater cartridge in the print head is managed by a PID control loop. This will always have some overshoot and undershoot of temperature when the flow changes substantially. There are a number of contributing factors to this issue in the printer hardware design. I could go into a lot of depth here but that is an aside.

    I would not use thin floating sections. Let's say the whole backing was solid in a FreeCAD design body. I would then do something like an egg shaped ellipse pocket out of the middle. Another option might be a % like shape with the thinnest section offset so that the layer lines are still substantial.

    My most advance approach would be to print the face of the frame on the bed, and the rear face of the backing plate also on the bed as a second part. If designed well from the start, and if the bed is large enough, you design the print to finish the backing plate before the face is completed. Then you add a print pause, remove the backing plate, insert it into the front plate, and continue the print to encapsulate it as a single part. This makes any 2D pattern for the backplate possible, and you do not need to deal with fasteners or whatnot. You end up with a perfect picture frame slot using this method.

    A total aside, but this idea can also be used to make your own printed supports manually for perfect overhangs. You print the support to size, add a pause, and remove the printed support shape. Ideally, you add a ~0.1mm-0.2mm clearance gap, paying very close attention to how your slicer layers height and first layer correspond to the support dimensions, or rotate the support to utilize better x/y dimensional accuracy if possible in some designs. You can even create a printed alignment jig on the build plate just to hold this manually created print support. The trick is to then apply gluestick to the interface between the manually printed support and your overhang. This can produce nearly first layer like print quality on an overhang with dimensional accuracy too.

    Another super advanced trick: let's think if the picture frame standing vertically upright in Cartesian planes. It is facing forward on the X-Z plane (X = -><- = >< = left to right). Let's assume the origin 0,0,0 is properly centered in the frame. Now if we look at a X/Y Top, section view, we are looking at the picture frame as if someone had used a hacksaw in the middle of the sides of the frame. In other words, we are looking at the frame's profile view. Now typically, people approach this like a [. Now this takes a lot of practice, but it is possible to design a profile something like ɭ̅̅̅ ̅ ̅˻ ̅ ̷̅ – the print bed is ↓. If you design this just right, the left side is the frame and the right is designed with a small connection to the bed and an angle where this connection is close to the rest of the frame. The thin bridge overhang is going to contract and shrink towards the larger heat mass of the frame, especially because of the printed layers above the bridge. This contracting force will be set into the part like a spring, but will remain compressed due to bed adhesion. You may want to add a small first layer connection or inner skirt to hold this section in place firmly throughout the print. When the part is removed from the bed, the bridge spring will pull the right section back. This will create the slot for your picture and potentially a way of holding other types of backing, while not worrying about conforming to other types of manufacturing process constraints like a wood router or sheet metal profile.

    Sorry if my lack of eloquence, verbosity, or tone come across negative at all. I wish to be encouraging and am just nerding out.

  • I have a dream. A dream of kaliaude code.

    There are already models trained on reversing binaries too.

  • People take off their mask and become their true selves when they feel anonymous on the internet. It sucks to see how many people have no independent ethics or character depth. It is a hard lesson, but this is true of humans everywhere. These same people act completely differently to your face in person, but they are ultimately the ones that hurt you most. They are a minority. Use the opportunity to learn about them and observe them. Then you will be better at spotting them when they wear the mask of social norms and peer pressure.

    Humans are tribal animals. We cannot escape that, but we can minimize it through self awareness. The number of perspectives, opinions, and intelligences are large. Everyone is stupid to someone, and in different ways at different times. The only stupid any of us can fix is within ourselves. Only worry about things you are able to change. Everything else is a waste of time. Sorry you had a bad encounter with someone.

  • 4chanGPT

    • ®#!% y$+ :%@) €&@:?*($#
    • $@& ®+$&#!
    • ¢+#% &% #%@&π©?
    • (open weights)
  • In all honestly, just try to talk to people. It does not always work. Some fuckwits are mods for the narcissistic idiocy. Be the change you want to see, and then push your way into the cultural discussion.

    Like 90% of the time, sending a PM is more than adequate to curb behavior. Every mod action is harmful without exception, as proven by the prisoner's dilemma in game theory. Negative feedback is incapable of producing positive outcomes. Adding as little negative feedback as possible creates the best environment for everyone.

    If you want to be a mod, that means sorting out flags by reading into the situation from both sides. I will not tolerate bigotry or personal attacks of any kind. However, every individual has a right to all information, a right to skepticism, a right to error, and a right to protest in nonviolent forms aka the right to offend others. No one has a right to infringe on these same rights of others.

    My personal feelings are irrelevant as a mod. I will not take actions against someone if I am participating in the conversation. I will flag the issue and PM another mod. I try to remember to sign my mod log messages because I have nothing to hide and despise any coward that lacks full transparency from a position of authority. I am the janitor. I clean up the messes. I read into the details of every flag. I am just as likely to take actions against the person that flagged the issue. I usually contact both parties in a PM too.

    I have experienced many bad mods, but that is my catalyst. Everyone is valid, even if unpopular. Everyone has a right to error because that is a primary way that we learn. The only scope that I will bias heavily is any issue with diversity and especially any transgender negativity because the Fediverse is a safe heaven in that regard.

    Dichotomous logic is a sign of stupidity and something to avoid with self awareness. If you are unwilling to reach out and take as little action as required, you a doing it for the wrong reasons and ultimately harmful. Immediate emotional reactions are harmful. Think stuff through from all perspectives and give every benefit of the doubt as possible. Everyone has a right to have a bad day too. Don't let some halfwit reaction make their bad day worse even if you disagree strongly as a person. A good judge always separates their emotions from cognitive logic. Being a mod is not hard at all. One just has to care. The community is self regulating with flags that indicate when action is needed. Micromanagement is a mental disorder. The best communities are often those with inactive mods.

  • My favorite surprise language is very very Asian looking with a hillbilly southern accent and slang. Really any distinctly atypical phenotype mixed with redneck is fascinating. I can easily opt out of my southern accent. It is not my default. I find it funny how some people cannot switch it off.

  • Full documentation and second sourcing of all hardware.

    This restores the right of ownership and destroys the current dystopian nightmare world of lost citizenship and democracy. It is closely tied to google winning the right to digital slavery and the buying and selling of your digital person to exploit and manipulate you.

  • 5 years. $125k was just the legal fees to break even.

  • I agree, but the tool is not to blame and never has been. This is not unique. The same thing was said with the advent of digital photo editing. Fakes predate that when people made them with film. In each instance, it is the ethics of the people that are the problem. At the deepest abstract levels, if you play out all of the implications, there is no difference between a tool and the thoughts in your head. Blaming tools is always a problem and only an illusion of a solution.

    I have been able to edit photos like this since before AI, but have never done so. When I was a kid in school I drew stuff like this. So the problems is actually intent and social accountability more than anything else.

    I don't like that the issue boils down to this. I think it is ugly, but I think the alternative is far far worse. I think those that are pushing this issue are convenient idiots combined with people that are hell bent on a fascist authoritarian future that is truly horrendous. If posting dubious child nudes of me will stop fascism and WW3, what poses would you like me to strike?

  • Democratic freedom and democracy in practice.

  • It will culminate in revising "minor" to something scientific. Age is arbitrary and impossible for AI alignment to handle. Neoteny and the Tanner Scale is visually present and real. That fundamentally shifts culture in large ways because retention of childlike traits is neoteny, and the scientific definition of beauty. That is well outside the scope of Anglo cultural awareness, and will be hard for the Puritan backwardness to adjust. Like Riley Reid is so popular because of her neoteny. Even with actors like Tom Cruise, the reason people like 'short people' in media is the neotenous head to body proportions closer to 1 to 5. All cartoons are doing the same but exaggerated. "Minor" must be scientifically grounded better than the age when male children are large enough to carry the kit of grossly immortal mass murder orgies the halfwits call war.

    Arbitrary age is irrelevant in purely visual media. Attempting to somehow mass surveil everyone to enforce it is dystopian insanity. The only solution is to be more scientifically grounded. It is ugly, and a bit ick, but whatever, those are illogical emotions fighting real justice and must be suppressed. Most countries like Germany and France have lowered their ages for minors. The only scientifically relevant ages are puberty and cognitive maturity at 25. Make it 25 and you will be supporting your kids until 25 financially, and the military will fall apart from people with the mind to question orders and think for themselves in mass.

    Edit: To be clear, I do not want people fucking kids or what not, or exploiting them. I'm against authoritarian parenting by proxy with a state making up for terrible parents, but that is another thing entirely. I do not care about this personally or in justification of anything. It is simply an observation of inevitability as data expands and ages. Images in 100 years will still be around just the same. Their categorization must be derived from the pixels themselves. There is no other long term solution. How anyone feels about that is totally irrelevant stupidity and amounts to nothing more than stupid children putting their head in the sand and saying "nuh uhh!" Science and democracy are ugly. They always have been. If you cannot handle the ugly with the good, you are a fascist or authoritarian, and to that dystopian nightmare culture, fuck you for starting a war for democracy soon. Sincerely.

  • Ask Lemmy @lemmy.world

    What are the alternatives to commodity housing?

  • Lemmy Shitpost @lemmy.world

    Let's go girls

  • Ask Lemmy @lemmy.world

    Is there a modern legislative alternative to "ignorance is not an excuse?"

  • Ask Lemmy @lemmy.world

    Admins of the threadiverse: What unintuitive oddities have you noticed in user behavioral patterns?

  • Programmer Humor @programming.dev

    0v3rcl0ck yu0r BR3AKFA5T!!!11!!

  • cats @lemmy.world

    dad joke landing

  • cats @lemmy.world

    Uber driver's genius plan to garner more tips

  • cats @lemmy.world

    Australian cats are different

  • Ask Lemmy @lemmy.world

    homeovestism - is it a thing? Preferred nomenclature? Respectfully correct?

  • Linux @programming.dev

    grep + tree how?

  • cats @lemmy.world

    skeptical

  • Ask Lemmy @lemmy.world

    Have you ever hired someone just because they were more attractive?