Here you go. A beautiful and open source news site article text scraper called “newswaffle” Now feel free as you browse the tomshardware articles with all that crap cut right out. I love it! Let me know if you are interested in how this works.
Its not just AI code but AI stuff in general.
It boils down to lemmy having a disproportionate amount of leftist liberal arts college student types. Thats just the reality of this platform.
Those types tend to see AI as a threat to their creative independent business. As well as feeling slighted that their data may have been used to train a model.
Its understandable why lots of people denounce AI out of fear, spite, or ignorance. Its hard to remain fair and open to new technology when its threatening your livelihood and its early foundations may have scraped your data non-consentually for training.
So you’ll see AI hate circle jerk post every couple days from angry people who want to poison models and cheer for the idea that its just trendy nonesense. Dont debate them. Dont argue. Just let them vent and move on with your day.
Thanks for sharing, knew him from some numberphile vids cool to see they have a mastadon account. Good to know that LLMs are crawling from “incompentent graduate” to “mediocre graduate”. Which basically means its already smarter than most people for many kinds of reasoning task.
I’m not a big fan of the way the guy speaks though, as is common for super intelligent academic types they have to use overly complicated wording to formally describe even the most basic opinions while mixing in hints of inflated ego and intellectual superiority. He should start experimenting with having o-1 as his editor and summarize his toots.
Me, a nerd: its just a joke don’t go off on a tangent, dont go off on a tangent, dont go - just let it be… I just… I just…
I NEEEED IT
So achtbually, nature works with transcendental real numbers on the complex plane with uncountably infinite precision.
Lets break this math nerd statement down in a way normal people might understand. First, most numbers are multi-dimensional and live on a plane instead of a line. The straight integer number line like 0, 1, 2 is just a small slice of the plane. In this plane, imaginary numbers occupy their own dimension. Complex numbers which are made of both real and imaginary parts occupy another dimension.
Moreover, most numbers are also infinitely precise thus being uncalculatable. Their decimal places go on forever and ever without repeating or being representable with a ratio of integers. Its why we only have good approximations for pi instead of an exact pinpoint knowledge of it. There are methods to get closer and closer approximations but you need an infinite time frame to complete that unending process.
Theres actually somehow more uncountably real numbers than countable integers and ratios, even though they are both infinite. There are more decimal numbers between 0-1 than there are integers between 0-infinity. One form of infinity is literally bigger than another, and that bigger infinity is the one nature likes working with.
Moreover, most of our universal physical and mathematical constants are transcendental reals because nature gets a hard-on for baking multidimensional fractal holomorphic topology and complex nonlinear equations into its magical abstraction bullshit logic.
Theoretical physicist during the 20th century were VERY salty about finding complex and imaginary real numbers in their physical equations. Since it implies that complex numbers arent just imaginary tools of abstrction but somehow “real” and affects the universes physical machinery. Nonlinear dynamic equations put a bullet through the brains of classical scientific determinism. Thank you very much, chaos theory and entropy.
It’s not that we invented imaginary numbers, its that they were the missing piece to fully complete our understanding of algebra. With them, we finally graduated from cave man linear algebra, to discovering holomorphic dynamics which model the way natural systems actually work. After 2000 years of banging basic logical abstractions together to make a enough decent sparks of discovery for a real smoldering fire.
Computer processing power sure helped to visualize these higher dimensional topologies for our little monkey brains to process with our eyeballs in real time instead of just thinking about this stuff in the minds eye. I sure cant visualize a 4D hypercube let alone a 20ishD hyperstructure that AI image network picture forms brought down to three dimensions.
Really its a miracle that we have even a thin narrow portion of numbers we can compute, all our regular integers and ratios are islands distanced apart by an infinitely deep ocean.
In case you were wondering about the stuff in the image: Multidimensional AI activation map showing how and image AI organizes its knowledge on a neural network. Similar concepts or images are closer together.
3D mandelbrot set with the logistic map highlighted along its real number line axis. https://github.com/jonnyhyman/Chaos
minibrot zoom in
algae colony arranging itself into conjoined 2nd iteration sierpinski triangle, screenshot from a journey to the microcosmos video.
pascals algebraic triangle encoding the sierpinski triangle by if the number is even or odd (base/mod 2)
the dynamic map of where a pendulum will land if pulled upon by three magnets equally spaced given its initial starting spot. https://youtu.be/C5Jkgvw-Z6E
I have a bad habit of mixing personal bias into things when I get into a passionate writing fit and it sometimes comes off as pretentious dickery. I never set out to attack kagi users themselves even if I can now see now it did come across and those comments were unneeded. I was being a pretentious jerk with those comments and apologize to all you kagi users for my assholery.
I do think its a waste to spend money on a search engine, and that open source software instance maintainers could probably make use that $ better than another search engine startup. I am being honest with those personal opinions. But its not my place to judge those who decide they are in a well enough financial spot to pay $ for a service that adds precieved value to their life or where they decide to pay it to.
Its fustrating as a FOSS nerd to see so many people shill yet another subscription based service feeding money into another souless company that makes promises of protecting your data and not selling it to ad companies now but has no gaurentee of holding those promises over time. That’s how the subscription services get you once they have you, slowly changing promises and creeping in their money making bs but slowly enough to not be too jarring. Maybe I’m just disillusioned with things after being burned so many times. Best of luck to you though I hope it continues to be a valuable service to those willing to pay for it.
This is a copy/pasted message I wrote up on another thread. As long as there are people in the comments shilling kagi, I will shill my prefered engines. At least my suggestions will bring awareness to free as in freedom projects. I hope to god people paying 10$/month just to not get datacucked by search engines will also learn something and save their money.
SearX/SearXNG is a free and open source, highly customizable, and self-hostable meta search engine. SearX instances act as a middle man, they query other search engines for you, stripping all their spyware ad crap and never having your connection touch their servers. Of course you have to trust the SearX instance host with your query information, but again if you are that paranoid just self host.
I personally trust some foss loving sysadmin that host social services for free out of alturism, who also accepts hosting donations, whos server is located on the other side of the planet, with my query info over Google/Alphabet any day.
Its nice to be able to email and have a human conversation with your search engine provider thats just a knowlegable every day joe who genuinely believes in the project and freely dedicates their resources to it. Consider sending some cash their way to help with upkeep if you like the services they provide, they will probably appreciate and make use of that 10$ better than kagi.
Heres a list of all public searx instances, I personally prefer to use paulgo.io All SearX instances are configured different to index different engines. If one doesn’t seem to give good results try a few others.
Did I mention it has bangs like duckduckgo? If you really need google like for maps and buisness info just use !!g in the query
search.marginalia.nu is a completely novel search engine written and hosted by one dude that aims to prioritize indexing lighter websites little to no javascript as these tend to be personal websites and homepages that have poor SEO and the big search engines won’t index well. If you remember the internet of the early 2000s and want a nostalgia trip this ones for you. Its also open source and self-hostable
Finally, YaCy is another completely novel search engine that uses peer-to-peer technology to power a big webcrawler which prioritizes indexes based off user queries and feedback. Everyone can download yacy and devote a bit of their computing power to both run their own local instance and help out a collective search engine. Companies can also download yacy and use it to index their private intranets.
They have a public instance available through a web portal. To be upfront, YaCy is not a great search engine for what most people usually want, which is quick and relevant information within the first few clicks. But, it is an interesting use of technology and what a true honest-to-god community-operated search engine looks like untainted by SEO scores or corporate money-making shenanigans.
I hope this has been informative to those who believe theres only a few options to pick from, I know these options are so unknown to most people.
I am a part of the Gemini protocol community. Newswaffle is a service hosted on the Gemini protocol to render web pages as gemtext (a simplified variant of markdown). Newswaffle the web article scraper is developed by Acidus. Here is newswaffles github .
I’m not a developer for it however I am one of the few people on this planet who actively use it and have had many email conversations with the dev over the years. Some of my suggestions made it into their services like lowtechmagazine be added to main newswaffle page and the simple English Wikipedia being added to their wikipedia gemtext mirror.
The github you linked is actually to the portal.mozz.us which is a seperate project that let’s me share Gemini protocol stuff like newswaffle over the web with regular people who dont really know about or understand Gemini and the small net. portal.mozz.us is developed and hosted by Michael Lazar (Mozz)