Structure 002: Viral AI

Welcome to this week's issue of Structure. As always, one topic, with the goal of providing you some extra insight into the themes you see online. If you enjoyed feel free to subscribe below!

Humans have been speculating about the future for decades. We constantly dream of flying cars, smart houses, and personal assistants built into every piece of technology we own. Some of these theoretical technologies have slowly crept into our lives while some have left the bulk of humanity waiting for progress. When Siri was first introduced by Apple, it felt like artificial intelligence (AI) had arrived. Over a decade later, I think we can all agree that Siri is a far cry from the droids in Star Wars or Kit from Knight Rider. 

The past few months have brought a variety of impressive breakthroughs in artificial intelligence into the spotlight. Is "the future" finally here?

This week we’ll be discussing some background and considerations around artificial intelligence. Whenever AI is brought into the limelight it forces people to speculate on the future. Here are some things to consider so you can decide for yourself if the future is really here or not. 

OpenAI

Led by CEO Sam Altman, OpenAI is an artificial intelligence research focused company with the overarching goal of creating AGI (artificial general intelligence) and using it to assist humanity at scale. The company has been leading research into AI applications for years, but their recent models DALLE-2 and GPT-3.5 have really brought them into mainstream spotlight. 

DALLE-2 allows users to generate raw images from descriptive text prompts. Anything you can describe, it can generate. GPT-3.5 has been utilized in their Chat-GPT application. A generative  text model, it allows users to essentially have a “conversation” with the model and use prompts to gather information, create outputs, or simply converse. Both of these models have been made accessible to the public through easy to use web apps directly on the OpenAI site. 

AI has become such a popular topic even outside of the tech community in large part due to OpenAI making these models so easy for the average user to experiment with. Very rarely is there a tech release that will appeal to both the average person and experienced industry professionals. That appeal to the average person is what helps this become a mainstream topic instead of a “tech” topic. Any cutting edge research company could benefit from the type of exposure OpenAI has received recently. Creating a simple interface that allows the world to interact with a new technology is a great way to slowly introduce and build excitement around a product. Personally this is a trend I would love to see cutting edge technology researchers continue going forward (writes down note to do a case study on the exposure gained from these public releases). 

Creativity??

Stemming from the introduction of the two OpenAI models introduced above there has been a lot of discussion around how generative AI will impact our creativity. Models that can create stunning images or entire blog posts using only small user-submitted prompts are incredible to experiment with, but are they meant to replace creativity? Probably not. These models will make it incredibly easy to achieve a certain level of result, however they will also open up the door to never before seen outcomes. While it will be extremely easy for users to rely on these models and lose some of their originality in the process, that won’t be everyone. Others will see these models for what they are, tools meant to augment humanity. 

 The chances of generative art replacing “normal art” are very low. However, generative art allows anyone to visualize something they may not be able to create themselves. Whether it be a painting, a logo, or a tattoo idea. Generative art provides an interesting new medium for expression that humans can tap into. 

The GPT-3.5 model has already gained some popularity with college and high school students for its ability to write entire papers or discussion posts. Naturally, educators are concerned at the thought of AI models writing papers for their students. That will likely need to be addressed with an AI detector of some sort, as it would be an injustice on education if left unaddressed. However, the model has also gained popularity as people have begun to create their own personal writing assistants. An idea I find especially intriguing, using the GPT-3.5 model to help finish sentences, brainstorm ideas, and help with editing. A great example of utilizing these models as tools instead of replacements. 

AI and Business

Another layer to the discussion on AI has been the applications it will have to businesses across the globe. Users have already created web-apps from the OpenAI APIs and leveraged profits. Speculation centers on a sort of “arms race” where businesses will be forced to either implement AI into their tech stack or acquire an AI focused competitor to stay afloat. I think this is a decent description of the long term progress of AI, but in the near term this is not super accurate (in my personal *unqualified* opinion). 

As with any new technology there are a few considerations that businesses will need to consider: scale, cost, and impact. Will the product be useful immediately at whatever scale the company currently operates at? How much will it cost to implement and use continuously? Will the efficiencies gained be worth the overhauls and costs required to implement? A lot of companies will look in the mirror and answer “no” to those questions right now. AI focused startups are spawning to attack the incumbents using their technology, but it will likely be some time before we see these companies taking significant market share from the competitors. New technologies usually take some time to iterate through improvements before companies start to jump on board at scale. 

That said, the recent breakthroughs in AI are the result of decades of work. This technology has been developing for a long time and we are finally starting to see improvements past Siri or Amazon Alexa. There will likely be significant value for companies that implement this correctly, first. There is always value in being first, so in the next 5-15 years I think it is feasible to say we will see companies attempt implementation. However there will be failures and losses before we see true winners and none of this will come all at once. 

I think consumers will be the true winners in this process over time. As startups grow and companies experiment we will likely see some significant competition over consumers. Consumers will likely have their choice when it comes to price point, methodology, and end result as we see AI work across industries over time. A dreaded thought for business owners may be a benefit for the every day consumer. 

The hype cycles we see in technology today often convince people that “the future” is right here staring them in the face. Things are not typically so cut and dry. Technological advancements often creep their way into our lives. They don’t announce themselves at the door with the amount of impact that will follow. It is easy to look at the past and mark where the advancements were, but in real time the world  is slowly changing every day. With that in mind, “the future” may not quite be here, but the stage is certainly set for a decade of incredible growth. 

Cheers,

Braxton

Song of the Week

As a thank you for checking out this week's edition of Structure here's a song I've been listening to lately:

If you enjoyed, subscribe below and feel free to give me a shout on Twitter!