Susi O'Neill
@susioneill.bsky.social
Tech communicator, speaker, author, thereminist, early electronic music nerd.
Weekly tech & AI insights: https://rethinkhypecycle.substack.com/
Weekly tech & AI insights: https://rethinkhypecycle.substack.com/
www.bbc.com/future/artic... content creators talk about "musical festival" instead of ICE protests, but are social channels suppressing content with keywords? They claim not but leaks show TikTok banned 'ugly' and disabled creators. BBC explore Algospeak and algorithm suppression
The words you can't say on the internet
There's a secret list of words you can't say on social media – at least, that's what everyone seems to think.
www.bbc.com
November 26, 2025 at 9:21 AM
www.bbc.com/future/artic... content creators talk about "musical festival" instead of ICE protests, but are social channels suppressing content with keywords? They claim not but leaks show TikTok banned 'ugly' and disabled creators. BBC explore Algospeak and algorithm suppression
futurism.com/artificial-i... AI slop is literally toxic for home cooks, with inedible recipes and bots scraping legit cooking sites and serving up in an unsavory AI gruel. The internet AI centipede is eating itself
Thanksgiving Dinner Headed for Tragedy as Disastrous AI Recipes Devour Internet
Food bloggers are warning home cooks to be wary of AI-generated recipes that could turn this year's Thanksgiving dinner into a tragedy.
futurism.com
November 26, 2025 at 12:38 AM
futurism.com/artificial-i... AI slop is literally toxic for home cooks, with inedible recipes and bots scraping legit cooking sites and serving up in an unsavory AI gruel. The internet AI centipede is eating itself
www.thepoke.com/2025/11/24/t...
Turns out Twitter's most vocal Maga supporters are red, white and blue. In 🇷🇺 who'd have thunk it
Turns out Twitter's most vocal Maga supporters are red, white and blue. In 🇷🇺 who'd have thunk it
Twitter’s public locations feature was such a disaster for the Maga accounts pretending to be based in the US that it was switched off almost immediately – 22 favourite reactions
Elon Musk has changed many things since buying Twitter – such as taking away functionality and reach from people who won’t subscribe, making likes private, rendering the block feature practically usel...
www.thepoke.com
November 24, 2025 at 10:22 PM
www.thepoke.com/2025/11/24/t...
Turns out Twitter's most vocal Maga supporters are red, white and blue. In 🇷🇺 who'd have thunk it
Turns out Twitter's most vocal Maga supporters are red, white and blue. In 🇷🇺 who'd have thunk it
ChatGPT introduces a 'chat' layer to allow collaborative projects. From family projects to business projects, this is an interesting step forward and starts to position ChatGPT as a social rather than solo network. openai.com/index/group-...
Introducing group chats in ChatGPT
Collaborate with others, and ChatGPT, in the same conversation.
openai.com
November 23, 2025 at 9:15 PM
ChatGPT introduces a 'chat' layer to allow collaborative projects. From family projects to business projects, this is an interesting step forward and starts to position ChatGPT as a social rather than solo network. openai.com/index/group-...
www.theverge.com/tech/825382/... The music industry still works on increasingly outdated models of composition credits and mechanical royalties from sheet music days. So interesting that the major labels are inking deals with AI music tech to allow remixes and licensing of real artists into AI tools
The music industry is all in on AI
After a year of lawsuits, AI companies and music labels are playing nice.
www.theverge.com
November 23, 2025 at 10:49 AM
www.theverge.com/tech/825382/... The music industry still works on increasingly outdated models of composition credits and mechanical royalties from sheet music days. So interesting that the major labels are inking deals with AI music tech to allow remixes and licensing of real artists into AI tools
x.com/TheHumanoidH... Another home helper robot that *may* in future be able to do the dishes, thanks to remote trainers creating vast data sets of mechanisms of minuate or tasks wearing sensor gloves
The Humanoid Hub on X: "Sunday Robotics unveiled its home robot, Memo, a wheeled robot with two arms and pincerlike hands. Training method: Sunday pays remote workers to perform household tasks wearing gloves that resemble Memo's hands. The Mountain View-based company is focused on a full-stack https://t.co/nEn9dDJ3Lf" / X
Sunday Robotics unveiled its home robot, Memo, a wheeled robot with two arms and pincerlike hands. Training method: Sunday pays remote workers to perform household tasks wearing gloves that resemble Memo's hands. The Mountain View-based company is focused on a full-stack https://t.co/nEn9dDJ3Lf
x.com
November 23, 2025 at 10:42 AM
x.com/TheHumanoidH... Another home helper robot that *may* in future be able to do the dishes, thanks to remote trainers creating vast data sets of mechanisms of minuate or tasks wearing sensor gloves
www.gartner.com/en/newsroom/... Is it prediction season already? Apparently so. Gartner's predictions always ones to watch: supercomputers, AI security and domain specific LLMs top this AI and security heavy list.
Gartner Identifies the Top Strategic Technology Trends for 2026
Gartner, Inc., a business and technology insights company, today announced its list of top strategic technology trends that organizations need to explore in 2026. Analysts presented their findings dur...
www.gartner.com
November 23, 2025 at 10:33 AM
www.gartner.com/en/newsroom/... Is it prediction season already? Apparently so. Gartner's predictions always ones to watch: supercomputers, AI security and domain specific LLMs top this AI and security heavy list.
futurism.com/artificial-i... AI data centres are being referred to by hypesters as infrastructure. Now reports say it's less Collosus and more like digital lettuce rapidly wilting. Horror flashbacks of Liz Truss.😳
Top Economist Warns That AI Data Center Investments Are "Digital Lettuce" That's Already Starting to Wilt
economist and author David McWilliams told Fortune that AI hardware investments are "digital lettuce" that are going to go "off."
futurism.com
November 23, 2025 at 10:04 AM
futurism.com/artificial-i... AI data centres are being referred to by hypesters as infrastructure. Now reports say it's less Collosus and more like digital lettuce rapidly wilting. Horror flashbacks of Liz Truss.😳
jmail.world just kicking off day reading a cache of Epstein mails in a Gmail interface. Weirdly snoopy
Jmail, logged in as [email protected]
You are logged into [email protected], Jeffrey Epstein's email. Sourced from the November 2025 House Oversight Committee data release.
jmail.world
November 23, 2025 at 9:16 AM
jmail.world just kicking off day reading a cache of Epstein mails in a Gmail interface. Weirdly snoopy
shecancode.io/why-plain-la... At work we often use internal lingo and in jokes. But what if it alienates others and leads to misunderstandings? Plain English and more straightforward comms could make tech teams more inclusive says @lisariemers.bsky.social
Why plain language is the key to building inclusive tech teams
Lisa Riemers and Matisse Hamel-Nelis explore how using clear, accessible language helps reduce misunderstandings, supports neurodiverse and multilingual colleagues, and fosters belonging.
shecancode.io
November 23, 2025 at 8:55 AM
shecancode.io/why-plain-la... At work we often use internal lingo and in jokes. But what if it alienates others and leads to misunderstandings? Plain English and more straightforward comms could make tech teams more inclusive says @lisariemers.bsky.social
Agentic AI can be more reliable by breaking down projects into microtasks. Cognizant AI's research showed greater task completion accuracy, even when using more basic LLM model. www.linkedin.com/posts/schegl...
Cognizant AI Researchers Achieve 98.57% Accuracy on Tower of Hanoi Puzzle | Mikhail Shcheglov posted on the topic | LinkedIn
This AI research paper will blow your mind.
Researchers at Cognizant AI have decided to test a hypothesis:
𝘐𝘴 𝘵𝘩𝘦 𝘨𝘳𝘢𝘯𝘶𝘭𝘢𝘳𝘪𝘵𝘺 𝘰𝘧 𝘵𝘩𝘦 𝘵𝘢𝘴𝘬 𝘤𝘰𝘳𝘳𝘦𝘭𝘢𝘵𝘦𝘥 𝘸𝘪𝘵𝘩 𝘵𝘩𝘦 𝘦𝘳𝘳𝘰𝘳 𝘳𝘢𝘵𝘦?
So they broke down a comple...
www.linkedin.com
November 21, 2025 at 5:30 PM
Agentic AI can be more reliable by breaking down projects into microtasks. Cognizant AI's research showed greater task completion accuracy, even when using more basic LLM model. www.linkedin.com/posts/schegl...
I love the sound of a Slack clickity-clack in the morning... busysimulator.com
Busy Simulator
Pretend you're busy by playing a bunch of app notification sounds. Made by Brian Moore.
busysimulator.com
November 21, 2025 at 5:27 PM
I love the sound of a Slack clickity-clack in the morning... busysimulator.com
An AI jury found a teen accused not-guilty, after several rounds of deliberation in a mock test. Could AI be less bias and more accurate to the facts than people in managing justice? www.unc.edu/posts/2025/1...
AI jury finds teen not guilty | UNC-Chapel Hill
The experimental mock trial at the UNC School of Law raises profound questions about artificial intelligence’s role in criminal justice.
www.unc.edu
November 21, 2025 at 5:27 PM
An AI jury found a teen accused not-guilty, after several rounds of deliberation in a mock test. Could AI be less bias and more accurate to the facts than people in managing justice? www.unc.edu/posts/2025/1...
www.instagram.com/reel/DRTJ8yw... The epic rap battle: Altman vs Viv dissing his claims (by Alexander Samuel)
undefined on Instagram: "awsamuelMy custom #GPT has strong words for #openAi CEO Sam Altman. So Viv put them to music, in a human-vs-AI rap battle…with SERI…"
awsamuelMy custom #GPT has strong words for #openAi CEO Sam Altman. So Viv put them to music, in a human-vs-AI rap battle…with SERIOUS footnotes. Because Altman’s side of this argument comes from what he has said and written about the “glorious future” of AI, like his promise to make it “broadly beneficial”. “Beneficial for whom?” AI Viv asks.You go, Viv. Hear more on Me+Viv —linkinbio
www.instagram.com
November 21, 2025 at 9:05 AM
www.instagram.com/reel/DRTJ8yw... The epic rap battle: Altman vs Viv dissing his claims (by Alexander Samuel)
www.anthropic.com/news/politic... Anthropic researchers test how politically neutral leading LLMs are. It's own, Claude, scored well and surprisingly even Grok surfaced counter opinions and even handedness
Measuring political bias in Claude
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
November 21, 2025 at 8:23 AM
www.anthropic.com/news/politic... Anthropic researchers test how politically neutral leading LLMs are. It's own, Claude, scored well and surprisingly even Grok surfaced counter opinions and even handedness
"We're like a family here" - if your family inflicted on you water torture, Victorian factory working conditions and sold your soul to the devil.
“The Sam Altman-founded company behind the ‘Orb,’ a volleyball-sized metal sphere that scans irises to prove people's humanity, had a message for employees: work weekends and ignore anything outside your job — for the good of humanity.”
Sam Altman's eye-scanning Orb startup told workers not to care about anything outside work
Tools for Humanity CEO Alex Blania said that "if you should care about something else, and if you want something else, you should just not be here."
www.businessinsider.com
November 20, 2025 at 7:49 PM
"We're like a family here" - if your family inflicted on you water torture, Victorian factory working conditions and sold your soul to the devil.
Unexpected AI in the bagging area – The future may be bots popping down the shops.
🔁 Get my newsletter: Rethinking the Hype Cycle #21 rethinkhypecycle.substack.com/p/unexpected...
🔁 Get my newsletter: Rethinking the Hype Cycle #21 rethinkhypecycle.substack.com/p/unexpected...
Unexpected AI in the bagging area – Rethinking the Hype Cycle #21
The future may be bots popping down the shops
rethinkhypecycle.substack.com
November 20, 2025 at 7:39 PM
Unexpected AI in the bagging area – The future may be bots popping down the shops.
🔁 Get my newsletter: Rethinking the Hype Cycle #21 rethinkhypecycle.substack.com/p/unexpected...
🔁 Get my newsletter: Rethinking the Hype Cycle #21 rethinkhypecycle.substack.com/p/unexpected...
I do really love Google's Nano Banana model for image generation. But tonight's it's gone while and decided to personify a robot with a shopping trolley. Rather than creating the damn image of one. 🤖
November 20, 2025 at 6:15 PM
I do really love Google's Nano Banana model for image generation. But tonight's it's gone while and decided to personify a robot with a shopping trolley. Rather than creating the damn image of one. 🤖
"If dead internet theory posits that the internet will eventually become only bots, alive internet theory proclaims we will never let the open internet die." Like an over-developed surburb, the web shifted from charming to alarming. How can we bring it back? news.spencer.place/p/alive-inte...
alive internet theory
why the bots will never win and just hanging out online
news.spencer.place
November 19, 2025 at 9:10 PM
"If dead internet theory posits that the internet will eventually become only bots, alive internet theory proclaims we will never let the open internet die." Like an over-developed surburb, the web shifted from charming to alarming. How can we bring it back? news.spencer.place/p/alive-inte...
A woman marries her AI chatbot in Japan. Hosted by company that specialises in “2D character weddings with virtual or fictional figures." Not legal. But legally blonde. www.independent.co.uk/asia/japan/j...
Woman ‘weds’ AI persona she created on ChatGPT
The 32-year-old developed emotional bond with virtual companion after painful breakup
www.independent.co.uk
November 19, 2025 at 9:07 PM
A woman marries her AI chatbot in Japan. Hosted by company that specialises in “2D character weddings with virtual or fictional figures." Not legal. But legally blonde. www.independent.co.uk/asia/japan/j...
LLMs may run out of training data next year after they've subsumed the entire history of human knowledge text. What next? More learning may come from computer game data x.com/RoboPapers/s...
RoboPapers on X: "With enough data, robots and AI can learn “world models” that let them predict the results of their actions. These models are a way to learn how embodied AI agents can perform a wide variety of useful tasks — but they require a huge amount of data. The team at General Intuition https://t.co/OvXhGnKuTQ" / X
With enough data, robots and AI can learn “world models” that let them predict the results of their actions. These models are a way to learn how embodied AI agents can perform a wide variety of useful tasks — but they require a huge amount of data. The team at General Intuition https://t.co/OvXhGnKuTQ
x.com
November 19, 2025 at 9:03 PM
LLMs may run out of training data next year after they've subsumed the entire history of human knowledge text. What next? More learning may come from computer game data x.com/RoboPapers/s...
@carolecadwalla.bsky.social 's been following the money and more predictions on an epic AI bubble crash, in a week where Sam Altman throws a hissy fit after being asked how likely it is OpenAI can be profitable. broligarchy.substack.com/p/the-great-...
The Great AI Bubble
Yes, it's a bubble. And yes, it's going to burst.
broligarchy.substack.com
November 19, 2025 at 8:52 PM
@carolecadwalla.bsky.social 's been following the money and more predictions on an epic AI bubble crash, in a week where Sam Altman throws a hissy fit after being asked how likely it is OpenAI can be profitable. broligarchy.substack.com/p/the-great-...
AI bias: LLMs are good at measuring sentiment, but terrible at non-American sarcasm (if there is any other kind of sarcasm) and more likely to label African-American text as hateful. theconversation.com/are-you-joki...
‘Are you joking, mate?’ AI doesn’t get sarcasm in non-American varieties of English
A new study shows large language models find it much harder to understand the nuances of Indian, British and Australian English.
theconversation.com
November 19, 2025 at 8:48 PM
AI bias: LLMs are good at measuring sentiment, but terrible at non-American sarcasm (if there is any other kind of sarcasm) and more likely to label African-American text as hateful. theconversation.com/are-you-joki...
www.politico.eu/article/brus... Changes to GDPR loom so EU can get ahead with AI. GDPR, but at the sacrifice of privacy measures. Could this be the end of the cookies pop-up? And the end of clicking accept on autopilot.
Brussels knifes privacy to feed the AI boom
Draft proposals obtained by POLITICO show EU is breaking sacred privacy regime to placate industry.
www.politico.eu
November 19, 2025 at 8:43 PM
www.politico.eu/article/brus... Changes to GDPR loom so EU can get ahead with AI. GDPR, but at the sacrifice of privacy measures. Could this be the end of the cookies pop-up? And the end of clicking accept on autopilot.
Google's Gemini 3 has landed! According to @emollick.bsky.social it's a notch up on reasoning, coding and organisation skills - getting closer to a PhD level student and showing AI models are still improving www.oneusefulthing.org/p/three-year...
Three Years from GPT-3 to Gemini 3
From chatbots to agents
www.oneusefulthing.org
November 19, 2025 at 8:39 PM
Google's Gemini 3 has landed! According to @emollick.bsky.social it's a notch up on reasoning, coding and organisation skills - getting closer to a PhD level student and showing AI models are still improving www.oneusefulthing.org/p/three-year...