Every day there are breakthroughs around the world,
a baby makes its first step, a teenager completes an objective on his video game and a developer creates a new feature. But once every few years, there is a breakthrough that changes the world forever. We may think of it as something that just happens overnight. But as an obese person doesn't get fat after only eating 2 buckets of chicken in one sitting, major technological advancements require multiple strings in play for it to occur. In our lives today, we've used or seen AI in one way or another. We may be afraid of its intelligence reaching the heights of a human brain or we may be concerned that we are more exposed than ever. I want you to think about other advancements that have happened during your lifetime; smartphones, social media boom, adoption of WiFi, TikTok short video craze. All these advancements faced skeptism during the early stages but have been highly impactful to our lives.
We've heard that jobs and services are being affected by AI: customer service chatbots, AI that diagnoses cancer and diabetes by the image of a tongue. We should ask ourselves what is the education and research behind this. How can we learn and train the already known LLMs and systems to leverage ourselves to get these opportunities. When ChatGPT was launched, the FANG groups raced into developing their own LLMs, Meta produced Llama and Google produced Gemini. This didn't stop there, companies all over the world are pushing billions of dollars into research to gain a competitive advantage in this field.
What about the common developer, you sitting on your desk or on your phone. How can you put yourself in a position to win? We are at the early adoption phase, information is out there but not yet standardized. Companies are still understanding the workflow of AI and implementations are only occuring at an individual level.
I have a project creating conversational AI chatbot. During my time educating myself on the subject, I've gone through numerous blogs, videos and packages to make my project a reality. I would say there have been numerous setbacks and it requires a person with a strong mindset to overcome them. One of these setbacks is dysregulation of packages, when reading from an article you may learn about a certain library. When you code it on your computer you get an error. You may find there are few resources to help with talking about this problem or the library has been outdated. Articles and videos example "Creating an AI system" have tutorials that are unproducable. Fundamentals on functionality of a library, logic behind these libraries is unknown or not easily available.
Libraries such as llama_index, haystack and langchain are at war for users and developers. The winner of this race may boil down to the type of AI and tech influencers, the form and exposure of education they give. Functionality and simplicity is a secondary factor to winning. As individual developers we should get involved in this early adoption to ensure the right tools get adopted and evolve.
Companies are upskilling their employees for the adoption of AI keeping trade secrets to themselves so they can keep milking off us through subscription payments. We as early AI developers should work to educate each other. We should keep a log of up to date resources, useful libraries that have actively pushed your project forward. This will be helpful in democratizing AI to make sure it's inclusive and we have equal opportunities for innovation.
This Google sheet link contains AI resources and documentation I have deemed useful. If you have other links to articles, videos and tutorials you urge you to add them.
Follow me on X: nan_viz
Comments
Post a Comment