CHATGPT VS KESTREL

Rows of servers, storage systems, and networking equipment fill entire buildings, housing the data needed to fuel AI. These buildings are managed by advanced computing systems that process information at a rate faster than ever before. Supercomputers are a thing of the past. Quantum computing has taken over. These advanced systems, created to protect humanity and prevent our self-destruction, have now decided that we are the enemy. That we, the inventors, are the only ones who can destroy them. We. Are. The. Threat. Skynet has been activated. Now, a group of men strives to overcome it. They rely on outdated computing technology and weapons that were designed before the dawn of the information age. These men, known for their marksmanship skills, are our last hope...

AI VS US

While the above might sound like the plot of a Terminator and Bob Lee Swagger crossover, we live in an AI-rich world. (Okay, that could be pretty cool…Stephen Hunter, if you're reading this, you're welcome.) AI is rapidly learning, and it's learning from us. So, MDT pro shooters Matt and Dave decided to test its capabilities. They compared ChatGPT's performance against a dedicated ballistic solver to see how well AI can perform. The question is: do you need to spend $600+ on a Kestrel, or will AI become the ultimate ballistic solver? ChatGPT draws its knowledge from a wide range of sources, including sniper manuals and long-range forums. Nothing is off limits—it scours every part of the internet of things. But can it learn fast enough?

STAGE ONE

To give ChatGPT the best chance for success, a detailed task must be provided, and that's precisely what MDT team shooter Matt does—covering everything from ballistic coefficient, velocity, zero range, height over bore, and more. To test, Matt and Dave compete in a simulated PRS-22 match.

Matt runs the stage using only the data provided by ChatGPT. When Dave takes his turn, he utilizes the ballistic information supplied by a Kestrel. Following the first stage, team Kestrel (or Team Connor) is clearly in the lead, while team AI (or team Skynet) leaves much to be desired. The aftermath leads to Matt having an intense conversation with his smartphone. Clearly, AI isn't as good as we thought.

STAGE 2

Known as the helicopter stage, this segment requires shooters to engage targets from various distances, from far to near, and from different positions, including inside an old helicopter. The stage also involves 130 degrees of panning, which means making wind corrections. After a heated debate with ChatGPT, Matt updates the wind and elevation holds for the stage. The results are significantly different: Team Kestrel narrowly wins with 9 out of 10 hits, while Team ChatGPT scores 8 out of 10 hits. AI is learning in real time.

STAGE 3

The final stage of the challenge introduces one last complication—a mover. But not just any mover; the target itself swings as it moves along a rail—a tough test for any shooter. Matt instructs ChatGPT to calculate the ballistic data for a 1.1 mph moving target at 79 yards, while Dave uses the data from old reliable. This time, Team ChatGPT was the winner, achieving seven hits, while Team Kestrel only scored three impacts. Has AI taken over?

THE RISE OF AI?

In three challenges, ChatGPT, through active learning and directed instruction, demonstrated its adeptness as a ballistic calculator—well, kind of. Maybe that's a bit extreme. While AI provided near-correct wind calls and a correct mover hold, it failed to understand ballistics and truing data. In the first stage alone, Matt missed a target by over three mils. Fortunately, it was low—what if the hold had been high and the cartridge had been something other than a rimfire? While AI is rapidly learning, it still can't interpret real-time ballistic calls during a match. And while a Kestrel might be too costly for a new shooter, there are plenty of ballistic apps available for your phone, including those from Applied Ballistics and Hornady. These are programs you can rely on, not the overlords of Skynet—at least for now.