Some Predictions For Next Year (2023) to 2030 (Basically A Note From Present Me To Future Me)

Some Predictions For Next Year (2023) to 2030 (Basically A Note From Present Me To Future Me).

As the title says, I am just going to list some things that I think may happen in the coming time. I just want to see how right the current AI tech bros are, including me. If it is really as exponential as we currently think.

Because as I watched some prediction videos from the past, we are not always right. And things actually take a lot of time to arrive. This is for 2023 and up to 2030.

Things that I think should arrive by 2030 at least. I am not being too optimistic as well. Like some people thought Proto-AGI will be made this year lol.

So future me, I know you are reading this. I hope this blog lasts in future, and doesn't get deleted for whatever reason because this is history right here that needs to be preserved. 

1. As Point-E Prototype was released and open sourced last night by OpenAI. This is basically DALL-E but for 3D model generation. It is not as impressive as DreamFusion by Google but 600 times faster than it. So I am expecting OpenAI to release Point-E to public which is much much better and Like DALL-E 2 currently is next year which is 2023. It may be named as Point-E 2.

2. Just like OpenAI releasing text to 3D model, I am expecting open sourced versions to release by Stability AI, creators of Stable Diffusion that is SOTA. And it changes industries like Stable Diffusion did. The quality I am expecting is something that will blow everyone's mind when compared to present, and will be usable on a consumer GPU.

3. Ofcourse other startups/companies will show or release them as well. So we will see the rise of text to 3D models. I see it being used in video games already by 2023. By indie developers or people who just wanna have fun. It will not be PERFECT but much better than current tech.

4. Text to Image models changed everything this year with DALL-E 2, Stable Diffusion and Midjourney. Artists went from this won't replace us to protesting on the entire Internet. This will continue next year with much better and advanced models. Expecting new models to release by other companies that will be easy to use like Stable Diffusion.

5. Emad said it will mature in 2023 so we will be able to do things that are currently impossible with it. Like basically replicating any type of art with incredible accuracy. Dreambooth which helps in training will get more advanced, faster to train and easier to use. It will replicate art very easily. Dreambooth alternatives will release according to me and by Stability AI.

6. Text to Image models will get much smaller in size, optimized and we will see them running on android mobile phones. It is already possible on iPhones. And I am talking about running it locally actually. So Stable Diffusion basically on everyone's pocket.

7. The speed at which it will generate will be so crazy, it will become real-time. It already is according to Emad. But on industrial GPUs. Distilled Stable Diffusion will release 2023, and will get more optimized over time. So real-time image generation will be possible on consumer level GPUs. Imagine generating images at instant. Maybe real-time on smartphones too. 

8. Because of that text to video models will emerge, we will see text to video AI releasing to public that is SOTA next year. And we can run it on consumer level GPUs. I don't know if we will be able to make literally movies with it by next year like Disney movies but by 2025 it may be possible. So anyone could make hour long movies by 2025 if they have the hardware. And it will be open sourced. So almost no restrictions as people would be able to train it like Dreambooth and Stable Diffusion.

9. Now VR/AR is my second most favourite tech after AI. I am expecting Meta to release Meta Quest 3 next year. Much better than Quest 2 but Quest Pro still better in some things.

10. Expecting Apple to finally show their VR/AR headset next year. It will be overpriced as usual but it may have some mind blowing features currently we don't see in the market. And people will adapt to it slowly. Making VR/AR more mainstream.

11. Expecting more companies and startups to release VR and AR devices.

12. Thanks to AI progress, by 2026 it may be possible to generate complete VR worlds just by using Artificial intelligence. And I'm not talking about Roblox like worlds but at least Gta San Andreas Level kind of worlds. In graphics and quality. It may be better even.

13. Expecting Unity and other game engines to add features that would make developing VR games much much easier. And anyone could create it with the required knowledge.

14. By 2030 I am expecting AGI to drop. And not single AGI but multiple AGIs by different corporations and startups. The world will never be the same

15. Because of the birth of AGI by 2030, the singularity will finally begin.

16. By 2030 I am also expecting that AIs will be capable enough to generate VR worlds just by seeing a photo or a group of photos and videos. So we could experience life of the past. What I mean is if you recorded a video of a place in 2022 and it has enough data, you could recreate it in future and live inside it. People in future won't just look at photos/videos as a memory but live it using VR.

17. GPT-4 should drop next year, before mid 2023 actually if rumors are true. I am expecting at least 50x performance over GPT-3.5. GPT-4 should be able to make conversations that are truly life like. So it is literally impossible to differentiate if it is a bot or a human. It will be so smart that it could come up with answers to any situation. So if you ask the bot, did you go to shopping recently, it will make up an entire story and convince you it did. And you won't be able to tell if it was a bot or human. 

18. I mean I could come up with so many things about GPT-4 but I don't wanna over expect and get disappointed later. But I see GPT-4 being at least used in video games to bring NPCs to life. Thanks to Whisper by OpenAI. We could talk to NPCs naturally. So it is basically Her movie coming to life, and everyone could get their own Samantha. The thing is OpenAI is very restricting, so idk.

19. ChatGPT is the current big thing going on. I am expecting open source alternatives to drop in 2023, one by Stability AI and I am expecting almost all restrictions being removed. I am not sure if they could match OpenAI's level though. 

20. In 2023 I am predicting that AI will go more mainstream, more people will use AIs in their regular life and talking to AI bots. More protests will go on by artists and other peeps against AI. 

and by 2030, we will look back at 2022 and laugh because everything is so much better now. Anything is possible, at least virtually. and after this, the world will be something no one can truly predict. But I hope it is good for us and not bad. 

That's it for now. Lets see how right I was, when I will read this in future. 

2022 was an amazing mind blowing year and I'll never forget this.