🤯 Top Microsoft AI Boss Concerned AI Causing Psychosis in Otherwise Healthy People
｢ Instead of coming up with meaningful guardrails, safety monitoring, or human counseling-based solutions, OpenAI gave in immediately, reinstating GPT-4o, and even announcing that GPT-5 would itself be made more sycophantic ｣
https://futurism.com/microsoft-ai-boss-concerned-ai-psychosis
🗣️ OpenAI Announces That It's Making GPT-5 More Sycophantic After User Backlash
｢ Those who had become accustomed to the "sycophantic" tone of GPT-4o, which sometimes lavished praise even on users' terrible ideas, were taken aback by GPT-5's "cold" brashness and short answers, highlighting just how emotionally attached many of them had become ｣
GPT-5 reaction amongst developers has been...mixed. In fact one of the developers OpenAI featured on its launch-day promotions has done a complete 180 — he now says he was wrong about GPT-5 (Theo Browne). Questions of coding quality aside, GPT-5 raises some interesting longer term questions for devs: if AI can build things just using web standards, will that lead to less reliance on React frameworks? https://thenewstack.io/gpt-5-a-choose-your-own-adventure-for-frontend-developers/ #gpt5 #frontend
openai put a deceptive graph... on the slide about how gpt-5 deceives users less than previous generations
the jokes write themselves at this point
#gpt5
openai put a deceptive graph... on the slide about how gpt-5 deceives users less than previous generations
the jokes write themselves at this point
#gpt5
"I believe GPT-5 is part of a larger process happening in generative AI — enshittification, Cory Doctorow’s term for when platforms start out burning money offering an unlimited, unguarded experience to attract their users, then degrade and move features to higher tiers as a means of draining the blood from users.
With the launch of GPT-5, OpenAI has fully committed to enshittifying its consumer and business subscription products, arbitrarily moving free users to a cheaper model and limiting their ability to generate images, and removing the ability to choose which model you use in its $20, $35 and “enterprise” subscriptions, moving any and all choice to its “team” and $200-a-month “pro” subscriptions.
OpenAI’s justification is an exercise in faux-altruism, framing “taking away all choice” as a “real-time router that quickly decides which [model] to use.” ChatGPT Plus and Team members now mostly have access to two models — GPT-5 and GPT-5-Thinking — down from the six they had before.
This distinction is quite significant. Where users once could get hundreds of messages a day on OpenAI’s o4-mini-high and o4-mini reasoning models, GPT-5 for ChatGPT Plus subscribers offers 200 reasoning (GPT-5-thinking) messages a week, with 80 GPT-5 messages every 3 hours which allow you to ask it to “think” about its answer, shoving you over to an undisclosed reasoning model. This may seem like a good deal, OpenAI is likely putting you on the cheapest model whenever it can in the name of “the best choice.”"
https://www.wheresyoured.at/the-enshittification-of-generative-ai/
"I believe GPT-5 is part of a larger process happening in generative AI — enshittification, Cory Doctorow’s term for when platforms start out burning money offering an unlimited, unguarded experience to attract their users, then degrade and move features to higher tiers as a means of draining the blood from users.
With the launch of GPT-5, OpenAI has fully committed to enshittifying its consumer and business subscription products, arbitrarily moving free users to a cheaper model and limiting their ability to generate images, and removing the ability to choose which model you use in its $20, $35 and “enterprise” subscriptions, moving any and all choice to its “team” and $200-a-month “pro” subscriptions.
OpenAI’s justification is an exercise in faux-altruism, framing “taking away all choice” as a “real-time router that quickly decides which [model] to use.” ChatGPT Plus and Team members now mostly have access to two models — GPT-5 and GPT-5-Thinking — down from the six they had before.
This distinction is quite significant. Where users once could get hundreds of messages a day on OpenAI’s o4-mini-high and o4-mini reasoning models, GPT-5 for ChatGPT Plus subscribers offers 200 reasoning (GPT-5-thinking) messages a week, with 80 GPT-5 messages every 3 hours which allow you to ask it to “think” about its answer, shoving you over to an undisclosed reasoning model. This may seem like a good deal, OpenAI is likely putting you on the cheapest model whenever it can in the name of “the best choice.”"
https://www.wheresyoured.at/the-enshittification-of-generative-ai/
Apple conferma: ChatGPT in Apple Intelligence userà GPT-5 con iOS 26, iPadOS 26 e macOS 26. 🍏 Miglioramenti vocali, video, coding e ragionamento avanzato. Attivabile da Impostazioni. #AppleIntelligence#GPT5#iOS26
🎈 GPT-5 Users Say It Seriously Sucks
｢ But many users say GPT-5 is far from the generational leap that its moniker would suggest. It's more of a mix of steps forward and steps back, prompting widespread speculation that OpenAI is trying to keep costs down. After all, running large language models is a notoriously energy-intensive — and environmentally destructive — process ｣
The sound of thousands of academics scrambling to rerun their experiments now that their papers about to be submitted (not to mention their previously published ones) have become unexpectedly obsolete.
"I speculated that transformer performance would converge on not-quite-good-enough. Needs more work. See me after. Not so much 'super-intelligence' as 'super-mediocrity'."
https://codemanship.wordpress.com/2025/01/11/the-llm-in-the-room/
The time is upon as, folks. If anyone doubted that LLMs have hit a performance wall, it's undeniable today. This is as good as they're gonna get, and it ain't good enough.
"I speculated that transformer performance would converge on not-quite-good-enough. Needs more work. See me after. Not so much 'super-intelligence' as 'super-mediocrity'."
https://codemanship.wordpress.com/2025/01/11/the-llm-in-the-room/
The time is upon as, folks. If anyone doubted that LLMs have hit a performance wall, it's undeniable today. This is as good as they're gonna get, and it ain't good enough.
When you vibe code the live presentation given to millions of viewers and don’t bother to check whether the y axis makes any sense whatsoever. Presumably the incremental increase from 69.1 to 74.9 just looked too small before? The AI equivalent of buying a flashy car? #GPT5
When you vibe code the live presentation given to millions of viewers and don’t bother to check whether the y axis makes any sense whatsoever. Presumably the incremental increase from 69.1 to 74.9 just looked too small before? The AI equivalent of buying a flashy car? #GPT5
A space for Bonfire maintainers and contributors to communicate
![Screenshot of Settings menu with "General" tab is selected on the left sidebar, which also includes options like Notifications, Personalization, Connectors, Schedules, Data controls, Security, and Account. On the right panel under "General," several settings are shown:
Theme: System
Accent color: Default
Language: Auto-detect
Spoken language: Auto-detect
Voice: Play
Show legacy models: toggle is turned on
Show follow up suggestions in chats: toggle is turned on
A large red arrow is pointing to the "Show legacy models" toggle.
[Alt text mostly written by ChatGPT, using Model 4o.]](https://cdn.masto.host/mastomachliscom/media_attachments/files/115/005/226/583/718/769/original/1483035e04096d19.jpeg)
