The future of 2013 is already here - My take on where we are headed

Recently I stumbled upon the movie “Her”. A thought-provoking film that was released in 2013, where a lonely writer develops an unexpected relationship with his newly purchased operating system that's designed to meet his every need. This exploration of love and the complexities of human connection in the digital age challenged my perceptions of intimacy and the boundaries between technology and emotion.

Not too long after that I came across a LinkedIN post from Dr. Joerg Storm, where he speaks about Lexi Love, an AI girlfriend created by FoxyAI, who has generated $360,000 a year by engaging fans who pay to interact with her, showcasing the booming market for virtual companionship. She offers 24/7 availability, language diversity, and personalized interactions. This phenomenon raises substantial questions. Within the same week, I watched a documentary on Swiss National TV regarding how entire agencies exist to hire ghostwriters that interact on behalf of models on OnlyFans. I have also heard demos of supposedly real-time AI customer calls that if they can scale it to an enterprise solution are very impressive. My point here is, “Her” in 2013 sure hit the right notes. How much longer until we use these types of technologies in agriculture?

Realistically probably a few years. I don’t think that the technology is the issue, I believe that it is the leadership in many organizations that is not ready to adopt it (I hope I can help alleviate that a little bit). I think that companies that decide to embrace new technology will have better margins and gain a competitive advantage, but that’s why you are here in the first place ;-).

How does the future look in 2033?

Let me start with this, my brain does not have the capacity to make predictions of where we will be in 2033, in a world where major announcements from AI companies occur weekly, however there are a few themes that I would like to mention.

Search

I read an article by Doug Clinton, who examines the impact of AI on the future of search engines, highlighting Google's significant earnings from its search function over the past five years and questioning whether AI innovations by competitors could threaten its dominance, namely OpenAI and it’s Bing search capabilities. He outlines the 4 reasons why people search and talks about where most of the value can be created.

It seems to me that we are in this amazing time, where currently many questions can be answered through ChatGPT. How many more years until OpenAI will monetize search as well? How do you as a company prepare for that? Let’s make an example of a search.

Which glyphosate should I use. Provide a brand

While ChatGPT provides you a brand name (based on millions of datapoints and content) Google provides you the brands that pay to be on top.

Two things that I would recommend taking note of: It pays to have a content strategy in the long run (P.S. With LLM’s this has just become much easier as well) to potentially rank in LLM’s and secondly it begs the question as to when and if you will see paid ads in ChatGPT.

The physical product you will be addicted to

This to me is an interesting esoteric question, especially through the lense of different learner types. Learner types can be broadly categorised into four main styles: Visual, Auditory, Reading/Writing, and Kinesthetic. Visual learners absorb information best through visual means such as diagrams, charts, and videos. Auditory learners prefer listening to explanations and benefit from discussions and lectures. Reading/Writing learners excel through interaction with text, enjoying reading and writing as methods of learning. Kinesthetic learners thrive on hands-on experiences, learning effectively through doing and experimenting.

The thing that stands out in the movie “HER”, is that he mainly interacts with the device via audio, so this type of technology would work well for auditory learners. This sure seems like the direction that the world currently is heading. Take the Hu.ma.ne ai pin or the Rabbit R1 as examples of audio uses of AI. However, voice solutions can situationally pose challenges - think sensitive info, coworkers, noisy environments etc.

Technology that would be more beneficial for the visual learner types encompass Apple Vision PRO and Brilliant Labs. I have yet to see solutions that are built for writer or kinesthetic learner types. (BTW, for a writer type, do they learn as efficiently when typing on a keyboard vs writing by hand? Let me know if you know the answer).

The holy grail is probably that you can just think of something and the task will be carried out. There are many research projects that show promising results by reading brain waves, so we are probably close to that reality as well.

However, I do believe that the next major step will be a combination of both visual and auditory combined. I therefore believe that apps will have to create ways that you can also interact with them via audio. Take Uber for example - why do I have to go into the app, enter my address, confirm the ride, when instead my device could be intelligent enough to make those tasks by itself with a simple voice command. “Hi Uber, I want to go home”. With the GPS location and knowing your home it will just carry out all of the tasks required. You would think that Siri from Apple is moving towards that direction. Which leads me to my biggest unsolved question that I have been asking myself these days. Is there a future for any apps?

What will be the one-stop shop?

The one thing that is clear to me is that we need to consolidate the chaos that comes with having hundreds of apps, each with their specific function. There is nothing efficient and streamlined with having 40 tabs open on your browser, or scrolling through 60 apps to find the right one. Dropbox vision therefore very much resonates with me. “Her” also has one place, one interface, one single source of truth. ChatGPT Memory is moving into that direction as well.

There are three options for me as to who could win that race:

1. Your Cloud Provider

   Which AI solution you will use in the future, will ultimately be driven by the cloud service that your company or you are signed up for. AI is fast becoming part of a vertically-integrated proposition. Microsoft Azure is quickly rolling out their Co-pilot suite of products (Sales and Service co-pilots have just dropped) - whilst Google is integrating the service fast in to their Gmail and Google desktop products.

   The open secret of AI economics has been that the real money was made from persuading corporates to move their data stack from on-premise servers to the cloud. Getting "ready for AI" was an attractive wrapper for what is pretty much a real-estate play. Hence players like Amazon Web Services (AWS) offered ever more sophisticated sets of ML tools to attract, and lock in, corporate clients.

2. Your preferred LLM with personal RAG

   Never in human history have we seen adoption of a software as fast as OpenAI. While their current attempt at bringing everything to one place with the GPT Marketplace is a bit clunky, I would not want to underestimate their abilities to create the one-stop shop. Layer in with that their memory features that they recently announced.

3. Apple

   They absolutely own how we interact with technology physically. Imagine a Siri that is 100x more powerful than it is today built into your glasses or lenses.

From a agriculture business perspective, I am asking myself, what does that mean for a future company? Will my customers in the future access parts of a SLM or LLM that we will have built internally? If so, webpages will become obsolete. How do we make sure that my company will surface in the one-stop shop? What content do you need to create to be served up? I have more questions than answer :-D

Stock Image Prompts in Midjourney

Below are prompts you can use to create stock images specifically for agriculture. Could you tell that they are AI generated? There are definitely subtle errors, such as the growth stage of the corn plants which don’t quite represent V3 growth stage.

Video just changed for ever

This week marked a significant milestone in video technology with OpenAI unveiling "Sora," their latest innovation in video creation (see below). The progress achieved within just one year (see what it looked like 1 year ago) is astonishing, prompting speculation about the future advancements we might witness in the next three years. Currently, the effectiveness of AI in this domain largely depends on the user's skill in crafting prompts and their creativity in envisioning desired outcomes. This development heralds an era filled with both excitement and challenges.