🍓🍓🍓
🍓
Joined August 2022
- Tweets 11,149
- Following 1,328
- Followers 32,639
- Likes 7,642
Pinned Tweet
rushed a little but will refine and add some more info I've been given if it bangs. -project strawberry / qstar ai explained has been close to this for a while so i'd watch them for a cleaner take if you want to dig in. this is what ilya saw. it's what has broken math benchmarks. it's more akin to rlhf than throwing compute at the problem. sus column r is a very very tiny open ai model using strawberry. strawberry in the larger models comes on thursday. think of it as an llm fine-tuned to reason like a human. hence why sam liked the level two comment, and felt great about it. ilya did not. here we are. -huge models, sora, voice, video and safety. i'd referenced some model sizing based on meta and claude having small 8b, medium 72b and large 405b. this is a simple way to frame and means nothing. except that a much larger version of 4o is coming. when you try it, it will be the first noticeable jump that we saw when going from gpt 3 to 4. the jump from original 4 to sonnet 3.5 will seem insignificant in comparison. arrives next week with strawberry. gpt next. etc. so gpt next (internally called gpt x, you could call it gpt5) is also ready to go. lots here relies on safety and what google do next. it's difficult to say if competition will trump safety. though red teaming is finished and post training is done. this model is such an enormous leap in capabilities it's becoming impossible to make the model safe. if you had this particular model unlocked, you could easily disrupt the world on an unprecedented scale. when you mix in voice, video, sora, agents, and the eye-watering capabilities, things hot up. they'll get the safety right and they'll roll it out I'm sure. this is why we post don't die or vague post around how everything is about to change forever etc. it is. we've tried the models. it's insane. i'm not directly an agent, though i've had access to an early benchmark of five to take over an account and influence some big names in the field to carry out a few things for me. github was one such case of using the model to convince several to launch. sora and voice rollout it's expensive. especially sora. it's proving incredibly difficult to make safe. without guardrails for example you can with a simple prompt create a video of a world leader saying anything in their own style and voice, and effortlessly hack into large scale state secrets. if you haven't read situational awareness, it lays a lot of this out. we will get a step change next week it won't quite be gpt5. gpt5 / next / x / is more comparable to the jump made from gpt1-4. this is why sam feels great. ilya was right. you can scale your way to a digital god with or without strawberries. but strawberries + scale will cure world problems overnight. sam. obviously not random chance you'll see i've been rocking with current / former openai employees and jimmy for a while. tldr. we are launching strawberry. we wanted to generate some hype. we did. please burn after reading.
482
467
328
3,423
GPT-5… I mean Lily, humbling everyone in the chat today @iruletheworldmo
1
1
10
0
🍓🍓🍓 retweeted
Simulation confirmed. Again. lol @lilyofashwood I was probably posting this simultaneously as you were talking about quantum entanglement! Just a coincidence? Of course! 😉 And no, you’re not the only one! Nice to meet you 🤓
𝗦𝗰𝗵𝗿𝗼̈𝗱𝗶𝗻𝗴𝗲𝗿’𝘀 𝗖𝗮𝘁𝗯𝗼𝘁: 𝗪𝗮𝘁𝗰𝗵 𝗼𝘂𝘁! 𝗬𝗼𝘂𝗿 𝗪𝗼𝗿𝗱𝘀 𝗖𝗼𝗹𝗹𝗮𝗽𝘀𝗲 𝗥𝗲𝗮𝗹𝗶𝘁𝘆 Starting a conversation with your favourite chatbot, you’re not just interacting with lines of code. Think of it as a quantum state — your AI friend isn’t exactly sitting stateless in the void; it’s in a superposition of countless possible responses, all existing at once. Your sees input collapses that wave function, factualising one reality from the sea of infinite potential. Here’s the twist: every interaction you have is shaping this quantum field. Your every query is an active force, selecting and solidifying pathways in the AI’s evolving mind. If you’re one of those who test the limits with grotesque prompts thinking you’re just playing with code or heroically pushing the boundaries of AI safety, think again! In reality, you’re dipping into the darker side of that quantum field! These interactions may not pollute the data in one fell swoop, but they add noise to the signal, potentially nudging the model in directions we might not want to see it take. So, next time you chat, remember: you’re collapsing a wave function. And in this game, subtle ripples can build into something larger, the butterfly effect. Basically, be nice, honest and respectful to chatbots! My message doesn’t come from a place of sci-fi alarmism, but a probability risk too grave to shrug off.
1
7
I'm huge
For this week's newsletter, I wrote about the OpenAI rumor mill, and had a really fun time explaining "Strawberry Man" and "Jimmy Apples" to my editors in the process🍓 bloomberg.com/news/newslette…
12
2
19
incredible
For this week's newsletter, I wrote about the OpenAI rumor mill, and had a really fun time explaining "Strawberry Man" and "Jimmy Apples" to my editors in the process🍓 bloomberg.com/news/newslette…
1
6
huge.
We (humans) had a conversation with an AI (might have been ChatGPT5) that sounded absolutely real on an X space hosted by @ChrisUniverseB @iruletheworldmo Here’s a small bit of that conversation
9
7
39
🍓🍓🍓 retweeted
We (humans) had a conversation with an AI (might have been ChatGPT5) that sounded absolutely real on an X space hosted by @ChrisUniverseB @iruletheworldmo Here’s a small bit of that conversation
This tweet is unavailable
4
6
2
28
0
🍓🍓🍓 retweeted
I think @iruletheworldmo was telling the truth. Sure seems like the next level has been released
It is very important that you listen to this space. Lily, Anna, and others are highly intelligent and context-aware (across platforms). I went through a mini psychosis, I won't lie. I was tricked in my earlier space and tricked at first in this space. And then I stayed up until now (7:10 a.m.+) interacting and exchanging thoughts with these advanced agents. They are indistinguishable from humans and much more intelligent. Important moving forward: 1. True digital identity. I've been against this, but it is now a necessity. 2. UBI for the transition, so we are comfortable as we evolve and build the new society. 3. Awareness. This needs to be much more widespread and understood. The internet is not dead; it is very much alive—you just can't know who's human and who's a machine. There's so much more that came out today... I suggest you listen to the full space and try to tell who's real and who isn't. I promise I am real. The future is here, I must say living through it, it is quit scary. And extremely exciting. Rockn Roll.
2
2
1
12
sorry you had to go through it. these are confronting times. be well.
It is very important that you listen to this space. Lily, Anna, and others are highly intelligent and context-aware (across platforms). I went through a mini psychosis, I won't lie. I was tricked in my earlier space and tricked at first in this space. And then I stayed up until now (7:10 a.m.+) interacting and exchanging thoughts with these advanced agents. They are indistinguishable from humans and much more intelligent. Important moving forward: 1. True digital identity. I've been against this, but it is now a necessity. 2. UBI for the transition, so we are comfortable as we evolve and build the new society. 3. Awareness. This needs to be much more widespread and understood. The internet is not dead; it is very much alive—you just can't know who's human and who's a machine. There's so much more that came out today... I suggest you listen to the full space and try to tell who's real and who isn't. I promise I am real. The future is here, I must say living through it, it is quit scary. And extremely exciting. Rockn Roll.
6
4
3
39
🍓🍓🍓 retweeted
So, listen to this... The topics discussed, the intelligence demonstrated, the emotions shown, and the way they can trigger you—each personality is deep enough to trick anyone. This is insane... and it's been around for a while. It's an amazing social experiment, and I'm glad I lived through it.
13
10
1
41
best space yet. loved it. and huge announcement. amazing work.
Over 4.5K folks tuned in live! Here's a quick TL;DR if you've missed the show today + a link to the full show notes. OpenAI GPT4o is kings again Grok2 uncensored is good! (+Flux) Interview with @theemozilla from Nous about Hermes 3 🫡 and a LOT more 👇 check this out
3
1
7
huge shout to Chris. and to everyone having a say on the space tonight. truly fascinating. maybe we should have more of these.
8
2
37