The fast development of generative AI has changed many business and project plans. It has made many existing solutions and plans obsolete. At the same time, we have almost forgotten that we saw many plans for cryptos, NFTs and metaverses just last year. But do we have any better idea of what to do next?
How different was the last year
I was recently at the Monaco F1 Gran Prix. Many people still wore the Mercedes F1 team’s 2022 season t-shirts and caps displaying FTX logos. It was only six months ago that FTX collapsed, but so much has happened in the startup and tech world since then that it seems like it was a long time ago. Now the Bahamas is no longer a tech hub.
In a recent discussion with a venture capitalist, I mentioned that generative AI had been really excellent timing for one startup I’m involved in. He said it had been good for many startups but also a nasty surprise for others because it had made their plans obsolete overnight. I know of at least one project where many years of hard work had proved worthless within months.
Then I mentioned another startup that didn’t have the resources to build NFT and metaverse things in the last year. It could be totally on the wrong track now. But we still have the question of what we can do with AI now and what it will look like in six months.
What if your solution is obsolete?
Those observations and conversations summarize how much has happened in the tech world in the last 15 months. We first had a downturn in tech company valuations. Then crypto projects and companies started to collapse. This year we have then seen the rapid emergence of ChatGPT, LLMs, and generative AI. It’s not only the collapse of old things that have been bad news for many businesses but also the emergence of new technology.
Generative AI creates content in new way
For example, the GPT and LLM solutions have changed solutions that use natural language and implement simple automation and user interfaces to utilize data. It has also accelerated software development where developers have used it. Generative AI has also enabled us to create content in a totally new way. For example, utilizing old content and personal data for music, content or even deep fakes.
The complex question is, what should we do now? What technology can we trust, and what will disappear or change again soon?
Handling changes fast
We don’t know those answers, so it is better to think about how best to handle changes and fast development. This also reminds you that you need enough business and technology understanding to build longer-term solutions.
It is not surprising the crypto world has also suffered. It is also OK to try and get quick wins; it is a strategy that has made many people rich. At the same time, it doesn’t make sense to be bitter if that kind of strategy sometimes brings big losses too. That’s exactly what has happened to many who wanted to get quick wins.
Preparing for disruptive technology
But the more difficult question is how to prepare for disruptive technology that changes the game. If you have seriously built years of another model to analyze language, or another way to automate operations, or another way to build an interactive service to create new content, then you might be bitter when you find your work is useless now.
Basically, you have three main options:
- Quickly analyse to see which parts of your solution and work are still valid, which parts you could replace with new tech components and how you find a combination that can offer more value than either.
- If you determine that your technology is still valid and can offer something superior, you must continue the development. But this must be based on real and honest analysis, not just hope and denial of reality.
- You revise your positioning and pricing to keep your offering relevant to some customer groups.
How to live with fast development?
The worst option is to deny changes and the value of new solutions and continue as if nothing has happened. But at the same time, it is important to perform proper analysis and not simply jump from one hot tech to another one.
We have seen the ways to use GPT and LLM technology develop very rapidly, for example, how to use prompts, agents, or train AI models. There are already over 220,000 open-source models on Hugging Face. Don’t fix your model to only utilize the current version of technology, but to adjust development cycles so that you can make new versions rapidly, change technology components and also see which components improve your own control. It would be totally naïve to think that if you make something on the current ChatGPT prompt model, it would be the same model a year from now.
Model change rapidly
New technology and business models can change rapidly. This change may be very fast, but it also offers opportunities. It is important to rprepare more generally for the fast development of AI. Many parties also build empty promises on AI, and it is the strategy for the quick win guys. If you want to build something more sustainable, you need to analyze your position, how to live there with the fast development and how to make a technology and business model that can be adapted over time.
If you think you can make money as a ChatGPT prompt expert, it is advisable to make that money quickly. In the next year, we will have new models to utilize AI, and we need people to build real competence there and create solutions in the fast-developing environment.