TrendPulse

THE FOLLOWING TEXT IS 100% HUMAN-SOURCED. TLDR: Those images I made are just parody ads of a near future where we could have an AI that sees and hears the same as us, and has enough context of our lives to decide what we do and how, in every situation. We could end up in a spot where all we do is select 'Proceed'. ------------------------- Today I was listening to the futurist Sinead Bovell, who was warning us about how we're increasingly outsourcing o...

Apparently dinosaurs make a come back

Since yesterday, the image generation is messed up. They always say that the image generation of is failed. Anybody experienced this?

https://preview.redd.it/c7i8d2t277eh1.png?width=822&format=png&auto=webp&s=5256b6b4056bd6fe0c7662eb444376d237351d48 its like this for more then 10 hours

I'm a small business owner who's been using ChatGPT, Claude, Canva AI and a few other AI tools over the last year. While they've been incredibly useful, I'm starting to feel like keeping up with AI is becoming almost as much work as using it. Every few weeks there seems to be a new model, new feature, or a better workflow, and I often wonder if I'm using the right tool or just sticking with what I already know. I'm curious whether this is just my experience or whether others feel the same....

What’s the nature of your chat? I have given mine a name and almost feel obligated to use manners 🫣. I know I sound crazy, but i feel invested in what i have created somehow. Please tell me I’m not alone on this? 🤣

so I installed the new ChatGPT desktop app today thinking it'd just be a normal update. instead I somehow ended up with two separate apps sitting on my Mac at the same time. one's called "ChatGPT (Classic)" and the other is called "ChatGPT Classic." not kidding, not a typo, go check for yourself. two apps, both named Classic, one just has parentheses. there is zero explanation anywhere of which one is actually the new one and which one is getting killed off eventually. I had ...

I have been reading about Inkling, the new open-weight model from Thinking Machines (from ex-OAI Mira Murati), and I am curious whether anyone here has spent real time chatting with it. The specs sound interesting. It is multimodal, has 975B total parameters with 41B active, and supports a context window of up to 1M tokens. But specs and benchmarks never really tell you what a model feels like in normal use. How is it for writing, coding, reasoning, and ordinary conversation? Does it have a ...

I asked it based on the other post that made the model lead itself to breaking its policy https://chatgpt.com/share/6a5c7230-7774-83e8-86d6-cc7bfe1a3946

https://chatgpt.com/share/6a5d1fe7-79d4-83eb-99af-74f51c9b23c6

This all started with a conversation about Spaceballs and if the Megamaid could fix the smoke problem in Canada that my government recently decided was something we should make Canada take care of.

I asked ChatGPT 5.6 to just clone a repo directly if it couldn't achieve something through the connector. At first, it said "oopsie that won't work, because can't reach the site from this container; I'll just use the connector instead" (I'm like, yeah, that makes sense actually), but some time later it goes like "I found a clean way around the container's DNS hiccup" and then it went ahead and used the hack to clone the repo directly.

I fell asleep with my phone open last night. It appears I clicked on the voice thing in my sleep. There is a LOT of this, but here’s a few highlights - including where it informed unconscious me that it is Human Resources 😭

Suppose it’s time to sort the laundry that’s been in the washing machine for 3 days. Edit: It's back up now. Y'all cant see the joke in replying "It's working what are you on about" to a post about the service being down at the point of OP.

Please write sequel to the movie Titanic (1997) where a Sith lord (Star Wars) travels back in time to resurrect Jack from the grave in 2026. The Sith lord convinces Jack if he becomes a Sith he will be able to resurrect Rose from the grave. Jack resurrects Rose after being his Sith training. They have a happy wedding, but the film ends with Jack realizing that he no longer wants to be a Sith. After Jack defeats the Sith in battle, the Sith decides to travel in tine somewhere else having lost aga...

Idk what they trained my boy on 😭 Full conversation: https://chatgpt.com/share/6a5c4c91-6b88-83e8-a54a-6f5c7bbc3513

I'm curious as to see how creative we all are at making AI sound as authentically human as possible.

I used the trial Plus version last time but ever since it ended weeks ago, I've strangely been still able to keep generating images with no limit hit. One day I even generated like over 20+ images in one shot. Before this, my image limit was always being hit by like max 5 a day or something. Sometimes even just once a day. Anyone else experiencing this or is it just me? Or did ChatGPT just increase the limit for everyone [link] ...

I don't get how that "fun alternative" wasn't blocked by the filters

Does anyone else still open Google first… …then immediately realize, “Nope, this is a ChatGPT question.” 😂

Built a small showcase site this week for a few real aircraft and rocket hardware, an F-22, an F-35, Starship, and the Raptor engine, styled to look like it belongs on SpaceX's own website. Saw someone else's version of this visual approach first and wanted to see how close I could get to it with a fully AI-driven pipeline. Everything started with three-view drawings in GPT Image 2, front, side, and back for each piece of hardware, clean enough that they could pass for actual technical referen...

English is not my native language. This post was translated into English by GPT. I was initially annoyed by ChatGPT’s new memory system. When it first appeared around two weeks ago, it could retrieve things from surprisingly old conversations, but the way it used them often felt unnatural. It reminded me of Gemini: the model would explicitly bring up old details when they weren’t really needed, almost like it was reading entries from a notebook just to prove that it remembered me. I dec...

Been trying to reproduce someone's interactive site this week, the one with those dense organic cell shapes that swim around and react as you move. Got the first two effects close enough that I wanted to write up how, even though the later ones are still running on placeholder versions. Step one was just handing the whole thing to GPT 5.6 Sol as a screen recording and letting it work backward: screenshotting frames, pulling out reference images, and rebuilding the site's structure and layout o...

Prompt engineering in 2035