Belfast man. Living in Korea, married to a Korean. AI nut. Talent/obsession for finding/ spotting new AI features. Have been mentioned in some websites? follow!
Joined March 2024
- Tweets4K
- Following985
- Followers997
- Likes564
Jimmy Carr actually said in an interview that young people shouldn't pay taxes until they're 25. Remember this? @jimmycarr theguardian.com/culture/2012…
Something needs to change for young people.
Watch the full episode on the High Performance YouTube channel: youtu.be/CotzNJNLEkE?si=-_Fr…
Hopefully Sora technology is doing similar things @OpenAI .
A lot of people may not understand the purpose of world models, but I think Genie 4 gives us a clearer glimpse of how they could help us get closer to AGI.
Google is already combining Genie with Gemini (unfortunately, 3.8 Flash in the live loop). Gemini can watch the simulation, break goals down into simpler steps and adjust its plan as the world changes.
Eventually, pairing an LLM’s reasoning with a world model’s learned physical dynamics could give agents a way to test actions and improve their spatiotemporal reasoning. Similar to how humans can imagine a real world scenario and think through what might happen before doing something.
This gives agents a place to compare actions and learn from simulated outcomes before trying them in the real world, assuming the simulations are accurate enough.
I think this could help with some of the spatial and physical reasoning problems models still struggle with. Humans can mentally rehearse what they’re about to do. But LLMs are only able to do text based reasoning and can describe physics and math possibilities, in text, but an interactive world model gives AI a way to play them out.
The physics and long-term consistency still need work, obviously, but this is a great step toward giving AI richer environments to interact with and learn from.
It's superb. Gemini isn't great in general. But, it's becoming incredible at world models. @GeminiApp @GoogleDeepMind
Merk retweeted
Here’s what’s new with Genie 4:
🔵 Photorealism: Built on Gemini Omni for lifelike detail across spaces and subjects
🔵 Advanced physics: Aerodynamics, collisions, and people who move naturally
🔵 Real-time controls: Prompt to change the scene, add elements, or interact with characters
🔵 Native audio: Ambient sound, music, background noise, and interactive dialogue
🔵 Gemini reasoning: Plans multi-step events to maintain coherence as scenes evolve
These capabilities are a step toward developing safe virtual environments to see how agents and robots navigate their surroundings.
We’re exploring early use cases with trusted testers first, before bringing the model to Google AI Ultra subscribers in Project Genie and early access partners in our API and @GoogleAIStudio soon.
Find out more → goo.gle/48idm3b
Merk retweeted
This is an event I had wanted to run for a long time.
I kept coming back to the idea, but never really knew how to approach it, so I kept postponing it. On September 28, we finally did it at @dev_korea: Accessibility, live, at AWS Korea in Gangnam, with a bit more than 80 people.
Most of us handle accessibility the same way: run the linter, check the contrast ratio, ship. Very few of us have ever watched someone actually use what we built with a screen reader. So we asked two blind and low vision professionals to show us, live.
👨💻 Inho Seo walked us through his day: KakaoTalk, maps, his IDE, SSH into a virtual machine, VS Code extensions, all by listening. I challenged him to go from his home screen to playing ANTIFRAGILE by LE SSERAFIM on YouTube. It took him about 20 seconds. And he had slowed his screen reader down so we could keep up.
🌐 Chan-Hong Kim couldn’t even reach a price filter on Coupang with a screen reader. With an agentic browser, the agent got it done. His takeaway: a more accessible web is a more agent-friendly web.
Of all our events, this is the one where I saw the most people really listening.
Full recap and photos 👇
dev-korea.com/blog/accessibi…
Merk retweeted
Y'all got it all wrong, I am making fun of OpenAI, and they deserve it.
They are being thrashed by Anthropic.
And they are releasing what, some blog posts and charging 100$+ for generous Astra usage.
Dario will release ASI on a 20$ plan next. They are sleeping. And people got the irony well.
You don't do this because you don't understand how bad the situation is.
A reminder that soon the $200 Pro plan becomes the same price, with HALF the useage. In no universe, is this not disgraceful, @thsottiaux @ChatGPT @OpenAI @AnthropicAI @claudeai
‼️REMINDER: ChatGPT’s $200 Pro usage gets halved on October 30 (Friday)
→ Codex and Work allowance drops from 20× to 10× Plus.
→ GPT-6 Pro: 200 → 100 messages a week.
→ Price: still $200/month.
→ Existing subscribers get credits that will expire December 31.
SO HALF THE USAGE FOR THE SAME $200/MONTH
Claude’s $200 plan is a much better deal…and you get Opus 5.5 which is the best Model at the moment…
Yet. They never do
"See you tomorrow for some good news."
I'm fine with it. I'm being patient, enjoying Claude's UI in the meantime, and waiting for Tibo's 28 Days to wrap up.
If OpenAI is smart, all they have to do is listen to what we've all been saying on X nonstop. We're literally the best source of feedback they could ask for. Just listen to us and thrive.
I've been getting dozens of messages from frustrated OpenAI users saying they were caught completely off guard by the Pro plan limits being cut from 20x to 10x.
And they're all expressing pretty much the same frustration:
"I'm already struggling with the current limits. Imagine having half of that."
"I was already thinking about switching to Claude anyway."
"They don't have the best model, and now they won't even have decent usage limits?"
And plenty of other complaints along the same lines.
I think this should be a wake-up call for the OpenAI team. They really need to pay attention, listen to this feedback, and use it to make things better.
Yes. Let's remember that ChatGPT Pro 20 is esssntially becoming GPT Pro 10 in 3 days.
Half the usage, for the same price. Absolutely disgusting, @thsottiaux No wonder people are mass exiting OpenAI products.
@ChatGPT @OpenAI @claudeai @AnthropicAI
ChatGPT Pro users, don't forget that your usage limits are getting cut in half in just 3 days.
Terrible news to be reminded of, of course. And the only way OpenAI can make this whole situation any less awful is by releasing a model that's as good as or better than Opus 5.5, while being way more efficient than the current GPT-6 Astra.
