• 36 Posts
  • 410 Comments
Joined 2 years ago
cake
Cake day: July 1st, 2023

help-circle





  • The lesson is that humans should always be held responsible for important decision making and to not rely on solely ML models as primary sources. Eating potentially dangerous mushrooms is a decision that you should only make if you’re absolutely sure it wont hurt you. So for research If you choose upload a picture to chatgpt and ask if its edible instead of taking the time to learn mycology, attend mushroom foraging group events, and read identification books, well thats on you.




  • Honestly, my favorite people are the ones who love to talk and are horribly desperate to babble to potential listeners. I’m not much of a talker but I absolutely dont mind looking you in the eyes and nodding my head as you talk about your hobby or current going ons.

    In bigger social groups I noticed this weird thing fellow humans tend to do where they all want a slice of being the talker/ center of attention and constantly cut off eachother or tune out current speaker waiting for them to shut up so they can start their monkey babble turn.

    This behavior absolutely infuriates me and I refuse to take part in it. I would rather just be silent and let you say your piece than interrupt the flow.

    As a knock on effect people subconsciously notice I’m not competing with them for talk time and am sending them constant listening signals like looking in the eye nodding head “mhm got you” stuff. This seems to really go a long way with making friendly with talkative types with minimal effort.


  • The question isnt whether quantum computers have an advantage over regular computers (they pretty much always do for code cracking as the parallel superposition computation is some crazy shit that changes cryptography forever) instead the question is whether or not AES-256 is able to resist our current quantum compute and how long it can do that.

    Its a simple equation, as long as it takes longer than the lifespan of the universe to compute with our most powerful supercomputers its considered good encryption. However as computers get more powerful, the projected time decreases potentially to the point of human lifespan time frames. Thats when it becomes a problem and the standard fails.

    Currently AES is quantum resistant but it almost certainly won’t be forever. New standards are gonna need to be adopted at some point.


  • Paid products can be enshittified. Also, its not just the quality of products that are getting enshittified but the concept of ownership over usage and access to digital data.

    • Slowly raising sub rates with that boiling frog tek.

    • No longer providing means to purchase local copies of data on a CD-ROM when you did before, just to pigeon-hole buyers down a subscription only access to the cloud.

    • Not offering a one time lifetime subscription in your sub-only model.

    It used to be that you bought something and owned it physically or at least owned a private copy of the data that could be cracked/ stripped of DRM so you could truly freely own and distribute. Now they all want to be digital landlords where you own nothing and pay a little more each month through the good old boiling frog while pinning price increases on inflation. The mid-term result is a 100$/year to rent out digital access to a dictionary when before you could buy a cd copy.

    Also, I don’t buy the “academic quality things should be incredibly expensive because its meant for scholars and university libraries” argument. Fuck that grift man. I know server infrastructure. It cost less to update a database or serve thousands of visitors than you might think especially for simple database lookups sent through https.

    It also cost practically nothing to distribute a digital file. So, Free digital access to educational and reference materials output by universities realistically should be a right in any sane society. Im sure Oxford University gets enough tax breaks and gov subsidy they could do it without impacting the stock holders precious quarterly figures. That entire 12 volume OED set + SOED takes up 500mb and can be fit on every modern tablet and phone. It sure as hell could be fit on a CD ROM years ago when they made that. The only reason its not is greed and maybe the dopamine rush scholars get from filtering the plebs.


  • so why all the fuss about the inaccessibility of OED?

    Because the OED is the creme of the crop for dictionaries, particularly the SOED has some of the most well put together definitions of any dictionary for casual lookup. Because the 1200$ paywall they put behind the physical editions was always bullshit. Because they no longer have legitimate ways of purchacing a cheaper local digital copy when one was available before is bullshit.

    Sure, wiktionary or webster might have an entry for the word but if you do side by side comparisions betweeen dictionary theyre mid compared to OED/SOED. If your reaching for one the logic should be that you want the best/most accurate and descriptive one possible, no?

    I genuinely believe that universities have at least a moral obligation (HA!) to provide free public services that better humanity. These are places of education subsidized and given tax breaks by the government for gods sake, yet theyre so corrupt from the rich fucks that run them like a for-profit corporation.

    I would make an argument that free access to the highest quality dictionaries thats the gold standard for scholarly reference and similar such materials should be closer to a digital right than anything. In a better world academia pricing structures get fucked, knowledge becomes truly open through digital online and local reference resources without DRM.

    Of course, thats a pipe dream. So instead, I simply ask for the option of an updated CD rom to be released as a possible purchacing option in a DRM free format. You know, like they already did years ago.













  • A fun weekend project was to set up a local model to tool call from openweather and wolfram alpha through their API for factual dataset retrieval and local weather info.

    Someone In our community showed off toolcalling articles on local instance of Wikipedia through a kiwix server and zim file and that seems really cool project too.

    I would like to scrape preprints from ArXiv and do basic rag with them. Also Try to find a way to have a local version of OEIS or see if theres an API to scrape.

    So I guess my solution is to use automation tools to automate data retrieval from wiki’s and databases directly. Use RSS, direct APIs, scrapers and tool calling.