A post by Rory Proud of Maritimedata.ai on May 26, 2023
I recently presented at the International Maritime Statistics Forum (http://www.imsf.info/) about the recent developments in AI, in part because I needed a sexy angle to address the fact that we still can’t agree on a universal vessel classification system (topic for another day), but mostly because I think it’s going to have a profound impact on the composition of our industry and the role data and analytics plays.
As illustrated by these timelines, the developments in AI have accelerated at a previously unfathomable pace.
Garry would probably appreciate we mentioning that he lost in the rematch, having previously taken the first leg 4-2 in Philadelphia a year earlier.
Humans have since defeated the best GoComputers. Kellin Pelrine, an American player recently beat a top ranked AI system 14-1 in a 15 game head-to-head.
I’m not sure it justifies another Terminator film, but I’m sure the Connor’s would be proud.
These are just some of the highlights that I’ve picked out this year to demonstrate the speed at which developments are happening.
It feels like a breakthrough release is coming every week, one of which that caught our eye was Open AI’s Chat GPT plug in, code interpreter.
This powerful plugin provides ChatGPT with a working Python interpreter in a sandboxed, firewalled execution environment, enabling the AI to handle uploads and downloads, solve mathematical problems, perform data analysis, and convert files between formats.
It allows users to complete complex programming tasks by simply asking a question in regular language.
It’s $20 a month…
So why bring all this up in a Newsletter dedicated to covering developments in the Maritime Data space?
Now people have access to the computational power available via these products, some of the biggest barriers to starting a competitive business (tech and infrastructure) has to a degree, been removed.
We’ve already seen a massive increase in the number of technology businesses all trying to address some of the more daunting challenges that our industry faces such as environmental and regulatory compliance, and for the most part that was before we could set supercomputers recursive tasks to execute while we sleep for less than 2 cappuccinos and a croissant..
GIGO becomes an even more important concept when considering that a lot of these models are similar in design and given the same inputs, produce similar results.
Therefore proprietary datasets that benefit from a degree of scarcity could be putting 0s at the end of their price tags as the demand differentiation grows.
It’s not all upside for Suppliers though.
Compliance could become a huge issue. As the number of micro-businesses without credible trading histories increases, using Bots to inundate Data and Solution Providers with requests for information and quotations, the effort to separate the wheat from the chaff could increase significantly, let alone dealing with breaches in contractual terms from business that could disappear as quickly as they’ve set up…
Humans historically are terrible at thinking exponentially and I’m no doubt massively underestimating the impact of this technology on our industry and the world. But one thing is for sure and that’s that interesting times lie ahead.
Commentaires