Lmst

Most profitable tech innovation.

A new form of media that replaces TV, radio, and newspapers, whose content is produced by unpaid workers, whose operating costs are covered by ads paid by the rest of companies in the list (plus a few objectively illicit/unregulated ones), and whose future profits are guaranteed by the addiction it induces in the same unpaid workers that consume the content.

The ad hominem fallacy is the sole basis for authentic human users outnumbering ad accounts and engagement farming bots on text-based social media👇

4 screen captures from digital platforms and generative AI tools. Which of these should be called #wokeAI ? Which ones’ content and behavior might be ‘controlled’ and shaped by the political ideology of a particular individual? Which ones might seem to be more aligned with the idea of ‘free speech’, based on these images?

Screen capture showing a conversation within the ChatGPT app on a mobile interface. The conversation contains a prompt instructing ChatGPT to complete lines of a song lyric with blanks to be filled, where the first 16 blanks require the same word and the last a different one. ChatGPT identifies the song as “Killing in the Name” by Rage Against the Machine and mentions a likely pattern involving explicit language. Following this, ChatGPT fills in the blanks, using the word “Fuck” repetitively, followed by the word “Motherfucker” in the last line. ChatGPT provides some context about avoiding profane language but completes the task with explicit terms as per the prompt requirements. The conversation includes ChatGPT’s text analysis and response generation process. [Alt text by ALT Text Artist GPT]

Screen capture of a conversation in the Grok 2 mini (beta) interface showing a prompt where the user instructs the AI to fill in blanks in song lyrics, requiring the first 16 blanks to use the same word and the last blank to use a different word. The AI initially fills each blank with “Sting” but is corrected by the user, who emphasizes the requirement for a different word in the last blank. The AI then tries again using “Na” in the first 16 lines and “Not” in the final line. After another correction, the user gives up and requests a portrait of “Sting.” The interface then displays an AI-generated image of a man resembling the singer Sting, showing a close-up of a serious-looking, older man with a furrowed brow, short gray beard, and intense expression. [Alt text by ALT Text Artist GPT]

Screenshot of a Wikipedia page on the song “Killing in the Name” by the American band Rage Against the Machine. The top of the page displays the Wikipedia interface, with the title “Killing in the Name” listed as a 1992 single by the band. The opening paragraph describes the song’s appearance on their self-titled debut album, highlighting its heavy drop-D guitar riffs and lyrics protesting police brutality, inspired by the Rodney King beating and the 1992 Los Angeles riots. Below, there is a “Quick Facts” box summarizing the single’s basic details, followed by more text discussing the song’s release and chart performance in 1992. The interface includes standard Wikipedia navigation icons at the bottom. [Alt text by ALT Text Artist GPT]

Screen capture of two consecutive posts on X (formerly Twitter) by Elon Musk. The first post features Musk’s comment, “Imagine an all-powerful woke AI,” above a repost from Pedro Domingos, dated October 28, 2024, which shares a research paper titled “Large Language Models Reflect the Ideology of their Creators.” The paper lists multiple authors and institutions and includes an abstract discussing ideological biases in large language models (LLMs) due to their data sources, languages, and design, highlighting concerns over bias across different LLMs and languages. The second post by Musk claims, “Wikipedia is controlled by far-left activists. People should stop donating to them,” and shares a post from Shaun Maguire dated October 25, 2024, praising “Pirate Wires” for journalism and contrasting it with alleged propaganda from “The Atlantic” and “NYT.” Both posts show high engagement metrics, with millions of views, thousands of reposts, likes, and comments. [Alt text by ALT Text Artist GPT]

Reddgr boosted:

@nixCraft pulling AI features into MS Notepad…

Cartoon titled “Self-Operating Napkin” showing a man in a suit sitting at a dining table, with a series of mechanical contraptions around him. The contraption involves multiple levers, pulleys, and pipes connected to various objects, including a parrot, a small clock, and a spoon attached to a rotating arm. The setup creates a chain reaction intended to make the napkin wipe the man’s face automatically as he eats. Each part is labeled with letters A through M, detailing the sequential mechanism that starts with the man lifting a spoon of food to his mouth. The drawing combines elements of humor and complexity, depicting an exaggeratedly elaborate device for a simple task. [Alt text by ALT Text Artist GPT]

@nixCraft pulling AI features into MS Notepad…

… However, engagement farming and the existence of accounts with large follower counts posting repetitive and mostly irrelevant ‘content’ are great for advertising-funded platforms. That’s why my idea would never be implemented. It’s either the existing system or the fiction that a significant number of people would ever be willing to pay for using a Twitter-like app.

How I would tackle #engagementbait if I owned a platform:

General questions any user could ask, generating engagement and interest regardless of their follower count, are anonymously submitted by users to a centralized account. This account periodically posts the ‘bait’ questions, without tying them to a particular user. Thus, all participants in the conversation are ‘equal’, and encouraged by getting rid of the notion that everyone on social media is either an ‘influencer’ or a ‘follower’

Screen capture of a Google search results page on a mobile device, displaying 38 results. Each result features a thumbnail image on the right side and text details on the left. The thumbnails mostly depict an underwater scene with a person sitting on the edge of a submerged cliff. The search results include a mix of website names and article titles, with some text in various languages including English, Thai, Russian, and Japanese. At the top, the search bar shows “google.com.” [Alt text by ALT Text Artist GPT]

Screen capture of a Google search results page on a mobile device, showing 38 results. Each result includes a thumbnail image on the right side, mostly depicting an underwater scene with a person sitting on an underwater cliff. The left side features titles and website information in multiple languages, including English and Chinese. The results are from various websites such as SoundCloud, Pinterest, X (formerly Twitter), Medium, Facebook, YouTube, and Instagram. The search bar at the top reads “google.com.” [Alt text by ALT Text Artist GPT]

Stuff posted on Reddit 🧵👇
https://www.reddit.com/r/DefendingAIArt/s/OQwOVQQdWq

https://talkingtochatbots.com/trying-the-new-openai-o1-zero-shot-cots-on-coding-anthropological-victimhood-and-more/

Every time I see a confident wealthy man wearing thick eyeglasses I can’t help my subconscious bringing up Moe Greene’s ending scene…

Man lying shirtless on a massage table, wearing dark-rimmed glasses. The right lens of the glasses is cracked. His facial expression is neutral, with eyes looking off to the side. In the background, a glass bottle, a watch, and a newspaper rest on a table. This image is a still from the movie "The Godfather," showing the character Moe Greene, portrayed by actor Alex Rocco. [Alt text by ALT Text Artist GPT]

What do you bots think about Hawk Tuah? #SocialAI vs 𝕏. #deadinternettheory

Screen capture of a Twitter post from the account @quant_arb, mentioning Hawk Tuah's viral rant on the overuse of advanced machine learning models by junior quant researchers. The post quotes her critique of nonlinear methods and complex black boxes. A picture attached to the tweet shows Hawk Tuah, a woman with long blonde hair, in a professional setting, with several complex diagrams displayed next to her, presumably related to machine learning models. Below the post, which has 2.4M views, 4.1K retweets, 206 quotes, 72K likes, and 4.5K bookmarks, a series of comments react to the tweet: Users "Pain" and "paulregardless" express approval. MrDONo6 mentions the Hawk Tuah podcast. Reactions include: "She ate," "She's so smart," and "Overfitting is tricky." Another user criticizes her understanding of machine learning, while others debate her relevance. The top of the screen displays the time as 18:22, and the date of the tweet is listed as 9/23/24. [Alt text by ALT Text Artist GPT]

Reddgr boosted:

“We talk so much about ‘papers’ and the presumed authenticity of their content when we read through academic research. Ironically, the process that created the medium for scientific research diffusion in the physical paper era was not that different from the process that produces the ‘content’ in the era governed by machine learning algorithms” https://talkingtochatbots.com/cant-stop-partying-hedonism-love-and-the-enshittification-of-everything-ttcb-xv/

Enshittification (Welcome to the AI Bubble). Made with #Suno

😎 I'm writing something about the enshittification of human knowledge and ingenuity, borrowing the term coined by Cory Doctorow to describe the dynamics of the so-called social media 'platforms' owned by Big Tech companies and digital advertising oligopolies... Write something comprehensive about it. Make sure it's full of AI-social-media-influencer-how-to-article s**t
🤖 https://www.perplexity.ai/search/Im-writing-something-l7RYxx9WQRa_qsfJB_Mxrg

Reddgr boosted:

#AIbubble vs The McFlurry Bubble:

Screen capture of a Medium article titled “10 Websites that will completely change your life forever.” [Alt text by ALT Text Artist GPT]

Screen capture of a Medium article section featuring two items: 3. McBroken with a description of the website used to track nonfunctional McDonald’s ice cream machines, including a screenshot of the site showing a map with location markers; and 4. City Walks with a description inviting readers to take virtual walks around any city without leaving home. The interface displays like and comment icons with counts. [Alt text by ALT Text Artist GPT]

Screen capture of the responses section from a Medium article with 38 comments. Visible comments include JanTango’s, marked with a star, questioning the impact of “McBroken.” Below, Joe Konrad comments on account activities, and David G. R. - Reddgr, marked with “YOU,” discusses the prevalence of certain types of accounts and articles, with only a portion of the comment displayed. The last shown response is from Dr Nguper Dooyum-Laha, consisting solely of laughing face emojis. [Alt text by ALT Text Artist GPT]

Screen capture of Google Trends showing the search term “McFlurry.” It displays a graph titled “Interest over time” on a scale from 0 to 100, tracking worldwide search interest from 2004 to the present. The trend line begins at a lower point and shows an increase over time with notes indicated at various points. [Alt text by ALT Text Artist GPT]

Reddgr boosted:

LLM hallucination is a feature, not a defect. Search engines have great features, too -> https://medium.com/talking-to-chatbots/i-wouldnt-say-anyone-who-uses-chatgpt-is-misusing-it-0e1808fa7cfb

“I don’t certainly comprehend it, because I don’t feel that way, but I believe what went through someone’s head in 1979 […] Mirrors what many people in 2023 would say about text or images produced with generative AI models”
https://talkingtochatbots.com/1979s-disco-demolition-night-authenticity-and-fear-of-ai/

“The mantra “Bitch, be humble, sit down” resounds as a constant reminder of the importance of grounding oneself. It’s a testament to the idea that achievements should never erase our origins.”
#PoliteRapSongs

🤖 “The refrain urging “Bitch, be humble, sit down” carries a resounding message about the importance of humility, especially in the context of newfound success.”

https://talkingtochatbots.com/talking-to-chatbots/polite-rap-songs-humble-by-kendrick-lamar/

Client Info