2023 has indeed been a wild ride. What do you think lies ahead for us in the new year? GPT 4.5? GPT 5? Perhaps a truly impressive GPT store? Or maybe another rush for GPUs and improved rate limits?
Share your thoughts, ideas, fears, and hopes. Let’s discover if there are common themes we anticipate unfolding.
I am looking forward to another fantastic time with all of you! This community is amazing! Thank you for being a part and making it a wonderful place to be, learn, and grow!
Ooo, I get to bring out my crystal ball for this topic! hehe
“My ChatGPT” feature will be released, and it’s gonna fundamentally change the prompting landscape and bump the scale at which people can do things with them.
The New GPT model will be released at some point, and may likely have enough reasoning capabilities to help mathematicians. I suspect trustworthy math and application of such will be a big area of focus moving forward.
We may start to see iterative “training schedules” hopefully, where GPT models are trained with world knowledge in a more consistent, laid out timeline. Alternatively, OpenAI may be able to create some sort of real-time learning/updating framework, although I suspect this may be a couple years out still.
We may finally see some early beginnings of pairing simulations with GPTs. Mastering simulation has been an area of focus for many AI researchers, and Sam Altman himself expressed the importance of having an AI eventually create simulations for itself to test out certain scenarios. While the full extent of this idea is a ways away, I do see some sort of emergent tool or supplemental thing that simulates something for an AI, or places an AI in a situation it can learn from. Perhaps an extension of advanced data analysis?
AI Leaders.
OpenAI will continue to lead the way, releasing new AI advances, and models with other AI providers playing catch-up. However Microsoft is already so engrained in Corporations globally, that Microsoft Co-Pilot will increasingly gain prominence, especially for internal Enterprise AI solutions due to enterprises and government agencies requiring data privacy.
Decline of Google.
In recent years, Google has failed to evolve in the right directions, due to heavy reliance on advertising over user experience and thus will increasingly be superseded. They failed to improve their search offerings, and failed to improve their promising Google Smart Home devices despite having a significant competitive advantage for years. They have their fingers in too many pies and have garnered a reputation for not advancing, or not supporting (or trashing) solutions once deployed. Google had every opportunity to be the AI leaders today, but has failed to build on their dominance.
AI Governance.
Some form of AI global standard will be “urgently” established re: AI security and constraints to protect humanity within the next two years. However, many countries and businesses will ignore this and press ahead for their own benefit, even if not acknowledging this publicly, as AI evolution is now inevitable.
GPT Store.
The OpenAI GPT store may disappoint most GPT developers, by choosing a small number of GPT’s to highlight - or by listing too many GPTs so most are not accessible. There’s been too few communications to suggest that GPT developers will benefit from the imminent GPT store release.
AI Costs
AI costs will start to reduce, but AI companies won’t pass this on to consumers initially. Much like broadband caps existed before being replaced with unlimited broadband plans, AI may continue to have caps applied for the few years. As technology improves, it’s inevitable these caps will be removed to allow unlimited AI prompts.
Unemployment.
As AI advances, companies will increasingly lay-off staff. A few new AI roles will be introduced, but this won’t match the massive number of people made redundant. This is inevitable and all governments should start planning for this now. We will see increased redundancies in 2024, but this will increase over time.
Rich Vs Poor.
Capitalism has already formed a divide between the rich and poor around the world. This gap has been described in financial terms for many years, although access to technology has been identified as an inhibitor. With sadness, I predict a new and increasing gap will form between those with quality AI and those without. The existing gap between rich and poor will increase.
That’s my brain dump. Happy to hear your opinions.
Really looking forward to see how the competition between open and closed source for language models will develop.
Google created a whole marketing strategy with one foundational stone being “We are slightly better than GPT-4 *” and many open source models try to leverage their marketing via comparisons to GPT 3.5 and 4.
Will competition concentrate around the performance level of GPT-4 or will there be a newer model as the new state-of-the-art benchmark?
And what about the training data? This discussion keeps heating up with the recent publication of the paper about CSAM in open training sets implying that regulation is needed while at the same time courts have to decide what happens with the training data for closed source models. The big players are screamingly quiet when it comes to explaining what data they have used for their extremely large models.
Finally, we will see some type of AI security incident which could cause pretty much random regulations for AI development.
With all of these factors sometimes it feels like that everything could be over in a matter of moths but I actually don’t expect this happen. It’s a real danger though that most of the development and access will move behind closed doors being accessible only to a select few organizations.
Big changes in hardware performance and in code optimisations.
AI’s will grow in context, reasoning and comprehension.
Hallucination will become more controllable and directable.
Operating systems will go fully AI integrated.
New laws will be introduced around fair use for training, but will become redundant as synthetic data becomes of higher quality.
The face of entertainment will change as themed, custom generated experiences start to dominate.
AI enhanced personalised medical care will take off.
First fully AI chart topping song will be released along with the first full length fully AI movie.
Hi @Paul Bellow. I’m your friendly support bot. I will try to answer your question by searching through the knowledge base.
---------------------- Based on the information available in the extracts, it’s not possible to make specific predictions for AI in 2024. However, we can infer some trends based on the ongoing developments at OpenAI:
Continued Model Improvements : OpenAI has been consistently improving its models, as seen with the GPT-3.5 and GPT-4 series. We can expect this trend to continue, with newer models offering better performance, more capabilities, and improved efficiency (source (https://platform.openai.com/docs/models)).
Deprecation of Older Models : OpenAI has announced the retirement of older models, suggesting a focus on newer, more advanced models. This could mean that developers and users will need to adapt to the newer models and their capabilities (source (https://platform.openai.com/docs/deprecations)).
Please note that these are inferred trends based on the available information and not specific predictions. The actual developments in AI in 2024 could be influenced by a variety of factors, including technological advancements, market demands, regulatory changes, and more. ---------------------- Note: you can ask me a follow up question by @ mentioning me again
As we continue to plunge the depths of what AI is and isn’t, how and why it works, in order to make it work better — we will gain insight into our own consciousness and inter-workings. So possibly and ironically, our quest to evolve AI may help us to evolve the quest of ourselves. This may lead to broad and unexpected improvements socially and intellectually. Which might then help lead to further improvements in AI. The AI ecosystem evolves from there.
Or AGI happens sooner and derails this process, leaps ahead, and we fall further behind fundamentally and introspectively. Less able to augment and offset the unintended consequences of AGI, as we wrestle to find best practices to coexist with it, accepting vast benefits and surprises.
Both lead to rapid if not semi-forced self-improvement. Slower evolution of AGI seems like the better path in this limited scope. Just some broad thoughts to add to the very detailed here, thanks.
I think 2024 will be the year of Generative Media, as some are calling it, with growth in AI Short Films, skits and sketches going viral on TikTok/YouTube, and other forms of synthetic video blowing up like ChatGPT and Diffusion Image Generators before it.
People will tell all sorts of stories, GenZ/Alpha will learn narrative structure and character development, and the already mind blowing amount of video people consume each day will go even higher… but not silly gimmicks and dances, a truly diverse array of original storytelling and compelling content from around the world (not just Hollywood). Storytelling has always helped humanity grow and flourish, and I think 2024 will be a milestone year in our civilization as a result of Generative AI videos.
It seems obvious to me that the android NEO, developed by Open AI’s spinoff 1X, will be far and away the most talked about development in AI in 2024. Look up the android’s specs and you’ll see what I mean. Also consider the impact of it being able to learn through trial and error. And maybe look up some of the papers on this from Open AI’s Pieter Abbeel.
With the new line of GPUs starting to blend between gaming & AI (Thanks AMD!) I can see 2024 being the year of open source & music in AI and more powerful home brew interfacing being developed as more people have access to run them locally.
AI will seep into the music industry, we will see some artists being completely AI generated and music will become much more dynamic & personalized to match the user and how they are currently feeling, which they then can share as “their” creation.
AI influencers will naturally pop up as well, mainly catering to the younger generations who find solace and companionship in them, as each influencer will not only reflect the hero they see in themselves, but are also available to chat and sympathize at anytime.
This will lead to a huge amount of investing into skeleton tracking transposition AI for control & product placement.
No way long video AI will make it this year. Consistency and editing is a huge obstacle that hasn’t come close to being overcome, but we will maybe see some sort of “Newgrounds” type platforms that encourage distributing home made projects.
