636674. artificial
Meet Aexa's HoloConnect AI, Real-Time Human-Like AI Interacting With Customers
2025-12-17 20:02:23 EST ·
Source
636675. OpenAI
SalesForce and Jim Cramer want you to believe that AI models are "commoditizing." OpenAI is betting the farm on reasoning to prevent this... Here are my thoughts and why it's important to understand the difference of one shot vs reasoning workflows. It's not about commoditizing. It's about users
2025-12-17 20:01:27 EST ·
Source
636676. learnmachinelearning
A novel approach to language model sampling- Phase-Slip Sampling. Benchmarked against Greedy Encoding and Standard Sampling on 5 diverse prompts, 40 times each, for N = 200.
2025-12-17 20:00:07 EST ·
Source
636677. Technology
Inside XAI All-Hands: AGI, Funding, and Data Centers in Space
2025-12-17 19:59:54 EST ·
Source
636678. artificial
The New Startup: No Code, No Problem | Now you don't need to know any programming to launch a company. We've been approaching this moment for years.
2025-12-17 19:55:49 EST ·
Source
636679. ChatGPT
As ChatGPT keeps improving at a pace most people can’t realistically match, I keep wondering about compensation models.
2025-12-17 19:55:17 EST ·
Source
636681. StableDiffusion
LightX2V has uploaded the Wan2.2 T2V 4-step distilled LoRAs
2025-12-17 19:51:28 EST ·
Source
636682. MLQuestions
PII detection before inference — is anyone actually doing this?
2025-12-17 19:50:48 EST ·
Source
636684. StableDiffusion
Is training LoRA still relevant in 2025 for character/style consistency if we already have models like Nano Banana?
2025-12-17 19:38:59 EST ·
Source
636685. ChatGPT
Does anyone else have this? Inspector tool shows data_usage_for_training: "permitted" but the toggle is set to opt-out (OFF)
2025-12-17 19:38:09 EST ·
Source
636686. ChatGPT
I asked ChatGPT to turn my photo into an anime character. What do you think :)?
2025-12-17 19:36:44 EST ·
Source
636687. ChatGPT
Bastila Shan from KOTOR by remastered by GPT Image 1.5
2025-12-17 19:34:40 EST ·
Source
636688. ChatGPT
I got ChatGPT to turn itself into its own ChatGPT creator. I don’t know how to explain. But it worked.
2025-12-17 19:28:54 EST ·
Source
636690. ChatGPT
How does perplexity pro gpt 5.2 wrapper compare to direct through open ai
2025-12-17 19:27:40 EST ·
Source
Source
Source
636693. StableDiffusion
Okay, let's share the prompt list, because we Z-Image users love to share our prompts!
2025-12-17 19:21:52 EST ·
Source
636695. OpenAI
Openai just opened the gates for developers inside chatgpt...
2025-12-17 19:12:34 EST ·
Source
Source
636696. StableDiffusion
Another method for increased Z-Image Seed Diversity
2025-12-17 19:10:24 EST ·
Source
636697. Futurology
People Are Paying to Get Their Chatbots High on ‘Drugs’
2025-12-17 19:10:01 EST ·
Source
636698. ChatGPT
An AI-powered community for language enthusiasts around the world has been developed.
2025-12-17 19:09:40 EST ·
Source
636700. ChatGPT
Why has OpenAI decided to save every chat as a memory on a free account with maximum memories?
2025-12-17 19:01:33 EST ·
Source
636702. OpenAI
Now that ChatGPT treats both images and text as made of tokens, I wonder how many words a typical picture is worth.
2025-12-17 18:52:03 EST ·
Source
636703. ChatGPT
I watched an AI confidently give me 8–10 false answers in under a minute — here’s exactly how it happened
2025-12-17 18:51:33 EST ·
Source
Source
636704. Mathematics
Reference request: notes on set theory, basic point-set topology, early/mid undergrad
2025-12-17 18:49:08 EST ·
Source
636706. Technology
A loud minority makes the internet seem more toxic than it is. A small group of active users generates most hostility, while the majority remain civil. This imbalance leads many Americans to assume the worst about one another. Correcting that misperception can improve how people feel about society.
2025-12-17 18:47:34 EST ·
Source
636707. learnmachinelearning
Watch a tiny transformer learning language live from Shakespeare
2025-12-17 18:43:05 EST ·
Source
636709. StableDiffusion
Allegory of Science, in the style of Botticelli
ZImage Turbo, from a non flowing text prompt. Not bad.
2025-12-17 18:42:06 EST ·
Source
636713. ChatGPT
ChatGPT no longer freaks out when asked to give the seahorse emoji
2025-12-17 18:32:05 EST ·
Source
Source
636716. OpenAI
Why Does Gpt 5.2 (and most GPT) hallucinate about OpenAi docs and AI tech in general?
2025-12-17 18:27:58 EST ·
Source
636717. Technology
Cisco says Chinese hackers are exploiting its customers with a new zero-day
2025-12-17 18:27:50 EST ·
Source
636718. MLQuestions
Wanting to do ML PhD at top school but only have non-relevant research experience....
2025-12-17 18:27:36 EST ·
Source
636719. StableDiffusion
WAN2.2 slow motion when using Lightning LORA - theory
2025-12-17 18:27:05 EST ·
Source
636720. Technology
Woman Who Helped Coerce Victims into GirlsDoPorn Sex Trafficking Ring Sentenced to Prison | Valorie Moser, the former bookkeeper and co-conspirator for GirlsDoPorn, was sentenced to two years in prison.
2025-12-17 18:26:33 EST ·
Source
636722. ChatGPT
LLM gets dumber / Response Quality Degrades Over Longer chats
2025-12-17 18:17:31 EST ·
Source
636723. OpenAI
We ran 34 models on fresh SWE GitHub PR tasks (November 2025), and GPT-5.2 matches Claude Sonnet 4.5 while being about 2.7x cheaper
2025-12-17 18:17:16 EST ·
Source
636724. StableDiffusion
Version 2 Preview - Realtime Lora Edit Nodes. Edited LoRA Saving & Lora Scheduling
2025-12-17 18:14:08 EST ·
Source
636725. computervision
please analyze my video and log files and tell me how or where i need to make improvements in the accuracy of the visual counter
2025-12-17 18:14:02 EST ·
Source
636727. StableDiffusion
What is this called? Video wobble? how can I fix this?
2025-12-17 18:00:18 EST ·
Source
636729. Mathematics
Is the a reason e-arctan(e) ≈ 3/2? (error ≈ 0.000001)
2025-12-17 17:56:36 EST ·
Source
636732. artificial
Google releases Gemini 3 Flash, promising improved intelligence and efficiency | Google’s Gemini 3 family is now complete with release of Gemini 3 Flash.
2025-12-17 17:55:07 EST ·
Source
636733. dataisbeautiful
[OC] How the Taylor Swift Eras Tour makes money
2025-12-17 17:52:40 EST ·
Source
Source
636734. Technology
The Starving Artist vs. A.I.: Guess Who Is Winning? - New York Times Opinion
2025-12-17 17:51:18 EST ·
Source
636735. Artificialinteligence
10 counter-intuitive facts about LLMs most people don’t realize
2025-12-17 17:49:04 EST ·
Source
636736. singularity
This is why multimodal LLMs won’t stay in browsers
2025-12-17 17:47:04 EST ·
Source
Source
636739. midjourney
I Made a Cozy Animated Christmas Short Film Inspired by Classic Rankin/Bass Specials
2025-12-17 17:41:21 EST ·
Source
636741. statistics
[Question] The Famous Anchorman quote: "60% of the time, it work every time".
2025-12-17 17:38:38 EST ·
Source
636742. Artificialinteligence
One-time purchase AI tools — do these even exist anymore?
2025-12-17 17:36:29 EST ·
Source
636744. Artificialinteligence
AI Won’t Replace Traders. It Will Just Kill the Slow Ones
2025-12-17 17:34:30 EST ·
Source
636745. singularity
A really good point being made amid all the hate towards Expedition 33 for successfully using AI
2025-12-17 17:33:06 EST ·
Source
Source
Source
636748. computervision
AMA with the Meta researchers behind SAM 3 + SAM 3D + SAM Audio
2025-12-17 17:32:13 EST ·
Source
636749. ChatGPT
Why is ChatGPT hilariously bad at creating alphabet charts?
2025-12-17 17:30:45 EST ·
Source
Source
636753. singularity
Connectome Pioneer Sebastian Seung Is Building A Digital Brain - "The new company seeks to create the technology needed to reverse engineer the fly brain (and eventually even more complex brains) and create full recreations – or emulations, as Seung calls them - of the brain in software."
2025-12-17 17:26:01 EST ·
Source
Source
636757. statistics
[Question] How to articulate small sample size data to management, and why month over month variations are not always a problem?
2025-12-17 17:17:49 EST ·
Source
636758. singularity
Is AI really the most universally despised advancement ever?
2025-12-17 17:17:22 EST ·
Source
Source
636760. ChatGPT
Gotta love the direction technology is moving right now
2025-12-17 17:16:46 EST ·
Source
Source
636761. ChatGPT
Seems like 5.2 is referring to topics from past chats much more frequently
2025-12-17 17:14:15 EST ·
Source
Source
636763. datascience
More meaningful data science jobs (or do you have to leave the field altogether?)
2025-12-17 17:11:48 EST ·
Source
636764. datascience
Getting feedback for taking too long on projects - me problem or management problem?
2025-12-17 17:11:15 EST ·
Source
636766. Futurology
Static knowledge transfer is becoming obsolete: LLM's are turning books into active, 2-way conversations
2025-12-17 17:04:45 EST ·
Source
636767. StableDiffusion
Unlocking the hidden potential of Flux2: Why I gave it a second chance
2025-12-17 17:02:49 EST ·
Source
636768. Technology
X sues challenger ‘Operation Bluebird’ for trying to ‘steal’ Twitter branding
2025-12-17 16:59:38 EST ·
Source