“She ( #LieslYearsley ) recalls an incident “many years ago” when she and her #Cognea co-founder #JohnZakos were in the office of a #SocialMedia giant, having built an #AmbientAI that could read a couple of sentences and get a pretty good read on someone’s personality, with a #conversational #UserInterface.”

“To my shame, I hesitated for a few seconds. I knew if I had said yes, we would have had an early exit and retired, but I knew this would lead to mass #disinformation. So I said ‘Yes, but that’s not why we are here’,” she says. “Someone else said ‘yes’, and we have the world we live in today.”

“Yearsley doesn’t regret walking away from the chance to make a fortune from #manipulating social media users, but it feeds into her cynical view of the motivations of the giant companies leading the AI charge.”

“They were there to pitch their technology as a way to detect young users at risk on the platform to provide support – but the unnamed social media giant’s head of mergers and acquisitions joined the meeting with different ideas. They asked if Yearsley’s AI could help it tell Republicans from Democrats, which it could, and also if it could be used to #predict what #news items #users would share.”

#Australia / #AIAwards / #AI / #ArtificialIntelligence / #technology / #Akin / #Startups / #founders <https://afr.com/technology/this-pioneer-wants-ai-to-be-a-force-for-good-20260526-p600xh?collection=p5zz3a>

AI Awards: AI pioneer and pessimist Liesl Yearsley wants to build AI that can show empathy and make humans’ lives better

Liesl Yearsley built successful bots a decade before it was cool. Now the trailblazer wants to build AI that makes humans’ lives better, not worse.

Australian Financial Review

The Haunted Hy-Vee Files

“The Haunted Hy-Vee Files” Strange signals and ghostly voices haunt a 24-hour grocery store. Jimmy uncovers a decades-old crime and a secret aisle that shouldn’t exist.

https://mybuddyjimmy.wordpress.com/2026/05/24/the-haunted-hy-vee-files/

#Google is testing new #conversational #adformats powered by #Gemini in #AIMode and #Search. These formats, including #ConversationalDiscovery ads and #HighlightedAnswers, aim to make #ads more #contextual and helpful by dynamically generating ad creative based on user queries. https://searchengineland.com/google-tests-new-conversational-ad-formats-in-ai-mode-and-search-478115?eicker.news #tech #media #news
#ChatGPT’s default model, #GPT55 Instant, is being updated to provide smarter, more accurate, and #personalised responses. The update includes improved factuality, a more natural #conversational tone, and better use of context from past interactions. https://openai.com/index/gpt-5-5-instant/?eicker.news #tech #media #news
GPT-5.5 Instant: smarter, clearer, and more personalized

GPT-5.5 Instant updates ChatGPT’s default model with smarter, more accurate answers, reduced hallucinations, and improved personalization controls.

OpenAI

New Blogroll Post

“Conversations all the Way” by Abhinav Tushar

«Full post on the site»

#Conversational-ai #Llm #Ml #blog #indieweb

https://lepisma.xyz/2026/05/04/conversations-all-the-way/index.html?ref=blr.indiewebclub.org

Conversations all the Way

Just when you thought #SocialMedia could *not* get any worse, this:

Zach Bussey 🇨🇦 -- @zachbussey
1h

#YouTube is testing a #Conversational #Search powered by #AI.

This experiment is available for #Premium members in the #USA.

Developer creates 'Conversational AI' that can run in 64kb of RAM on 1976 Zilog Z80 CPU-powered System — features a tiny Chatbot and a 20-Question guessing Game.

The venerable Zilog Z80 CPU has been around since 1976, and it has powered everything from calculators and home computers to arcade cabinets. But the 8-bit microprocessor isn't exactly a powerful CPU compared to what we use today. That said, developer HarryR has created Z80-μLM, a working "AI" for the well-respected microprocessor.

⁉️[HarryR] confirms that it won't pass the Turing test, but it is a bit of fun. And no, the price of Z80s will not be impacted by AI.⁉️

https://github.com/HarryR/z80ai

#zilog #z80 #conversational #ai #retrocomputing #engineering #media #retro #tech #news

Also started a small newsletter for my writings on #Conversational #AI which will be more regular from here on. Here it is https://buttondown.com/computational-conversations
Computational Conversations

Practical notes and reflections on Conversational AI systems

Just heard a take that "we’re all much more comfortable talking than clicking a bunch of buttons", with the conclusion being that we need to strive for building #conversational and #LLM powered interfaces for everything.
And, like, no? Except for a few specific situations, I very much do *not* want to talk to my computer if I can achieve the same goal by pressing a few buttons instead. Kinda makes perfect sense though that the people pushing hard for AI seem to generally be the, uhm, talkative type.

#AI

#OpenAI has upgraded #ChatGPT with #GPT51, introducing two new models: GPT-5.1 #Instant, which is faster and more #conversational, and GPT-5.1 #Thinking, which adapts its #reasoning power based on query complexity. The update also includes increased #personalisation options. OpenAI emphasises the importance of user feedback. https://venturebeat.com/ai/openai-reboots-chatgpt-experience-with-gpt-5-1-after-mixed-reviews-of-gpt-5?eicker.news #tech #media #news