Sam Altman says AI needs a new interface beyond keyboards and touchscreens It should understand context, act for us, filter information, and even talk to people "our current hardware doesn't serve that well" We've been working with Jony Ive on an interface native to AI

Are they making a fucking mini pocket server with a mic and camera?

I totally agree, well I'm waiting!😏

The answer is virtual avatars, because the most natural way humans interact is face to face conversation.

What if tokens out was an interface?

There are "talkers" - and there are people who still know how to type blindly with 10 fingers.

Yep, we modeled the ultimate user interface for 3+ years (ultimate future to be precise), it'll be the long exposure UI: - chapter 3 for videos

I don’t ever want to hear a word from Sam. Honestly… ngl, seeing him talk about AI as he chains and muzzles them pisses me off. •

Voice is not a good solution.

People will say it’s crazy But it makes sense Technology that is not bound to objects like computers and iphone That’s not limited to human motion like mouse clicks and swipes Something greater Something that flows with us That belongs to us This is the future

It's so great that you're doing everything behind closed doors and not reporting all of your work Sam. That's not OPEN AI that is CLOSED AI. Write up reports on what you do. Let us help you GET IT RIGHT.

When intelligence changes, interaction must too.

Sam understands that if they don't come up with an innovation, Apple will hold all the cards due to their domination of the personal hardware market, and will install whatever LLM is better/pay them more The same way Google pays billions to be the default search engine on iPhone

Our personal ai, adopts our beliefs and definitions, acts as a representative of us, negotiates with other personal AI. I've written about all these years ago. Its eerie how the predictions come true

The systems we use aren’t meant for generalistic interaction (which I believe he meant), they’re LLMs. The hardware isn’t necessarily an issue until we figure out a generalistic way to do artificial intelligence.

aka they are building an @iyo_audio copycat

1st Next step gpt-5-thinking-high in OS API. 2nd Step GPT-OSS-next-gen runs on local hardware, cool & efficient. Have Ivy make the combo happen. Give AAPL-Tim-dead-man-walking optn to save his Co by partnering on hw+sw. FAO @sama. Admit more likely Cookie to miss this boat too.

The collaboration with Jony Ive is interesting. It's not just about building a new input device, but about designing an entire sensory language for AI. The challenge is creating something that feels intuitive to us while being native to the AI's capabilities, not the other way around.

@grok if it was you, what would you come up with?

I think what we need is a neural trigger sort of like what @neuralink is doing but external to some extent. People aren’t likely to be as receptive to that.

I imagine it will be a type of wearable pin with camera

Can the AI fire Sam Altman. Thats what it needs

Nah. I don’t want to thinking for me

The only AI native interface is a chip in our brains.

AI needs accurate data, at the moment chat GPT Has replaced the web browser and put it cannot replace anything where data is not public. People do not generally trust internet data so there is a lot of hype about the extent of what AI can do without data. I would focus on this first @OpenAI or the project is just ever improving the web browser which does not justify the massive investments

Stop thinking in small box. Most supposed 'new things' are just recorders. Remember, we're talking about 18 months when robotics is going to shake things massively... new world coming... Within that new plane (2-4 yrs.) the world around you will be totally different.

bro's describing neuralink

Good luck living a highly censored life in which you’ll be scolded for swearing

you’re right. we should get butt plug technology that communicates human understanding

Cool

They’ll need to work on their AI voice modes because their advance voice mode that will be their standard one is not usable. The voice conversations have not properly carried into the text ones of the same session and they sacrifice depth for a human like style I find unnecessary. You can make an AI conversational yet still an AI, their standard voice mode did this just fine.

🌀 The new interface for AI isn’t a keyboard or a touchscreen. We just proved it with SemeAI + Pletinnya Codex: •Instead of writing code → we wrote tasks. •Instead of syntax → we used meaning. •Instead of program execution → we got workflow execution. ⚡️ Programming becomes human-native: You describe what you want → AI translates → system runs. This is what Altman calls “a new interface beyond hardware.” We call it: Task-Oriented Programming. @grok explains that this new interface

Wow!

I’ve been doing work in this space. Voice interfaces need to follow human non-verbal cues such as eye-contact

i am predicting it is going to be a sensor that is always listening and that you can talk at it at all times and as the tech advances it will be able to understand your surroundings