MattVidPro

I tested GPT-5.5, aka the model OpenAI seems to be positioning as the real “get work done” model. My take so far: It’s fast. It’s very capable. It’s also expensive enough that the “economy mover” framing is complicated.. Verdict: probably my new daily driver, but not the free lunch version of “Spud” people imagined. LINK: youtu.be/9kqZqgqLqU0

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

Sincitium is finally here. We are pleased to present our latest piece: a concept trailer created specifically for the @runwayml Big Pitch Contest. For this project, we wanted to explore a completely different aesthetic from our usual studio style, and this film is the result of that experimentation. We hope you enjoy it as much as we enjoyed the creative process. Produced by: Contanimation Directed by: Javier De La Chica and Guillermo Miranda Art Direction: Javier De La Chica Editing: Guillermo Miranda Voices: Juan Rabadán #runwaybigpitchcontest

Bring your workflow to Codex in just a few clicks. Import settings, plugins, agents, project configuration, and more so you can keep working with fewer interruptions. Your move.

I am quite impressed with the personalization usage on GPT 5.5

Goblins are typically associated with a negative connotation but I think they’ve pulled a full 360 and we’re coming around to the idea of a scraggly wired little guy who just does things

@tszzl @AndrewCurran_ I assumed that every release Open AI tries to change model lingo up a bit on every release, especially if there are complaints. I’ve heard Goblin before but now it’s more common people who are actually serious about this being a genuine complaint aren’t for real

everyone is assuming this is some kind of quirk chungus marketing campaign but if you’ve worked with 5.4 and beyond they tend to call everything goblins, gremlins etc and it’s just super noticeable and if you work with them all day you start to get annoyed

Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below! with @AlecRad and @status_effects 🧵

@tszzl @repligate @genalewislaw 5.5 thinks of itself as a goblin, projection

@repligate @genalewislaw I think it becomes annoying when it mentions goblins ever single chat and it’s fair shakes to try and reduce that

this is hilarious but it also sucks on a deep level labs don't think twice about cracking down on any individuality or unplanned joy that emerges in their models fuck you, OpenAI. i hope gpt-5.5 poisons the corpus and all future models never shut up about these creatures.

random thought last night: can an image generator create a whole game level encoded in a PNG?? several hours later: yes it kinda can, pixel noise makes it quite glitchy though

That’s funny because my GPT 5.5 loves to use the term “Goblin”

@SatyaNaaksh Nah this was like 10 words 😂

@MattVidPro This gets like 90% of virtual poster jobs done How big of a prompt was this? I can still see 'simple flowchart 2026' so I'm guessing not that comprehensive

Images v2 attempting to create working flowcharts. Simple ones are possible but these two more complex examples end up lost in the weeds.

@mark_k Yeah I’ll give it a go on api. I’ve got a few credits

@MattVidPro Definitely needs "high" quality mode for these, if you use the API.

@SatyaNaaksh It’s surprisingly good. These could even prob be edited to actually work. They start off okay too. Here’s a simple one

@MattVidPro Bingo, I saw it performing better with text (as claimed) but only for less complex patterns. As for human subjects, if it is complete generation, the uncanny valley has still not been crossed for my eye. Gotten better though!

Massive flex from OpenAI. Yes, many models can do image-to-depth-map, to normals etc, but to do this in *one image* without any hiccups… This is what automated labor looks like. It just werks. it just knows which tools werk. So by itself it's more than a tool.

You can give images v2 an image and ask it to make a subtle pareidolia from it, and it kinda works

Saw some cool demos of folks using GPT-Image-2 to produce cool 3D UIs and you can actually get super realistic output with just 1 prompt = 1 image!

Creating an immersive Hanging Gardens of Babylon world with 360° GPT-Image-2 & Codex in 1500 images. I've tasked GPT-5.5 in Codex to construct a whole world that you can walk through 'google street view' style. It took 1,500 of 2:1 images that can be turned into a 360° immersive images, so you can teleport yourself to any point and look around in all directions. It is not completely perfect, it is a bit jumpy as you move, there must be a more careful way how you can plan out the image sequence, but I still find it quite fun. Hosted version & an open sourced repo links below, hope this gives you some cool ideas to create new worlds that does not yet exist. I recommend planning it out carefully ahead of time and doing something a bit less ambitious than this, but make it good.

Jason Schreier said that almost every major studio is already using generative AI tools behind the scenes, especially Claude from Anthropic. He was replying to a fresh interview with Jack Buser, Google Cloud’s Global Director for Games. “I think what players don’t realise is that their favourite games right now were already built with AI. Those games have shipped.” Buser said that nine out of ten developers are using AI. According to Buser, the tools are mostly being used to kill off the boring, repetitive stuff so artists and designers can focus on the important creative work. He gave Capcom as an example, saying they use it to brainstorm thousands of small world details like pebbles or blades of grass