I just concluded a conversation with the Grok AI engine. Well, let’s just say I stepped away for a while. I can walk back in anytime. I decided to upload a copy of the Deer Log, and a couple of other supporting docs to Grok and see what Grok can figure out about our deer hunting history.

The Deer Log is a listing of every deer killed by our camp. It actually starts back in the mid-80s when I first started hunting. As my sons and friends joined me at camp, I added them in as well. As of this writing there are 99 entries in the log. I’m not going to post the whole log or list all the variables I track, but suffice it to say there’s a lot in there: the deer, the weather, the location, who shot it and what they shot it with.

Here is the conversation: (It’s about 34pages) . You can download it here: Deer Log Questions

I also sent up a listing of all our stands and blinds:

Site	Type	Location	Primary Wind
Faulty Towers	Ground Blind	Field	E
Ground	Ground Blind
Jagende Hutte	Ground Blind	Field	W
Lazy Boy	Ground Blind	Woods	W
Midway	Ground Blind	Field	SW
S10	Ground Blind	Field	E
Stump	Ground Blind
Blackberry	Tree or Tower	Field	W
Campground	Tree or Tower	Woods	SW
Fountain Square	Tree or Tower	Field	W
Garbage Pit	Tree or Tower	Woods	S
Heartbreak Ridge	Tree or Tower	Woods	W
Hollywood	Tree or Tower	Field	SW
Hundred Acres	Tree or Tower	Woods	N
Maple	Tree or Tower	Woods
Newstand	Tree or Tower	Woods	E
Virginia	Tree or Tower	Woods	N

. . . and this graphic of prevailing winds:

. . . a discussion of which is here:

Which Way the Wind Rose?

Discussion:

I’m flat-out amazed. I started deer hunting about the same time I started in Data Processing in the early 80s. I got in at the tail end of punch cards. Having a discussion with a machine like this is well-nigh magical. It read my spreadsheet, linked it together with another spreadsheet and an oddball graphic, interpreted all the data accurately, and made plenty of spot-on inferences.

Was it helpful? The sky did not part. There was no shaft of light from the clouds. However, Grok gave me a new perspective on things. It picked up from the raw data some basic truths:

1. The big bucks are not out in the pastures. They’re back in the woods. By laying off the treestands this year to let my legs recuperate, I’m giving up weight and antler for more deer sightings in general.
2. Mornings for us are the best time. I clued Grok into a bias in the data: if we bag a deer in the morning, we’re not going back out in the afternoon. We also only get one buck per season per law. Even with this, Grok says there’s a preponderance of morning kills. Interesting.
3. The deer log tracks kills, not trips afield. There is no data on how many times we went out and saw nothing. That would have been important for a lot of serious interpretation. It’s a bit late for our camp, 25 seasons into it. It’s a hint for y’all just starting.
4. We have enough good weather during rifle season, we often times don’t hunt in inclement weather. It shows in the data.
5. Wind is important, but not that important. The most kills are on a light southwest wind, but then that’s the prevailing wind.
6. Grok picked up a minor signal regarding moonphase. It says there’s a slightly better chance under a waxing or waning crescent. I did a study of this back in 2018 that did not see the 4th quarter as a particularly productive time. However, we’ve had several deer taken in the last quarter in the succeeding years. I still think a waxing moon is better than a waning moon and things shut down dramatically in the latter stages of the lunar month.
7. There are minor signals regarding weather. For instance, we generally kill bigger deer in colder weather. However, nothing to keep you on the couch.
8. Grok says that we’ve got a late Opener this year is good news or at least not bad news. I’d seen it differently.

Waaaay back, back in the early 80s, I got a hold of one of the early Chatbot programs called ELIZA. It was primitive, but it gave you somewhat of an illusion that you were talking with a human. I had a friend that got lonely and started conversing with this thing and got head-over-heals in love with the thing. Poor devil. He needed to get out more. Eventually, he started dating a U.C. Cheerleader and got straightened out. One thing that I notice about all these natural language-based systems is they tend to ingratiate, tend to confirm bias, and tell you what you already know. Grok is still that way, and I don’t think that sort of slavering is going away any time soon. However, once you get past that there is some real benefit to this sort of interaction with AI. As Patriarch of camp, I’ll continue feeding Grok information and seeing what comes back.

On the whole, it’s still still the whole Data Processor’s dilemma: GIGO (Garbage In/Garbage Out) . Where Grok wandered, it was because the data was incomplete, or it didn’t know about pre-existing biases. I tried there at the end of the conversation to have it make suggestions on where Moose, Angus, and I should hunt this year, and it stuck is in the same spot all season. That whole part of the conversation is kind of a irrelevant sidetrack, but it just shows the limits of this idea of Grok advising the patriarch of a deer camp.

One thing I will credit Grok for is its apparent knowledge of the mechanics of deer hunting and its ability to apply it to a bunch of data. It knew what was what and what it meant.

My advice to y’all is to read the transcript and see what you think.

Views: 9