BaseMap AI trained on Rokslide

Is this where autocorrect comes from?

Cause if it is, I don't know if I'm worried


I'm redneck and can't spell, but I'm convinced my autocorrect is more reforking stupid than I am.
Autocorrect has evolved a lot and the newer models are small versions of language models. They also do prediction now too. Smaller models can produce answers far faster than the large models but the tradeoff is often accuracy.
 
Well, if you notice it learns pattern prediction in language, not reasoning nor logic. Got some hilarious answers from ChatGPT when I asked it if an NF scope will survive a 100 foot drop onto concrete (it basically said it would).

Give them a question without a clear answer, and they are stumped.
All the research effort at the large AI companies in the last 18 months has focused on these deficiencies. I was working on some military projects 3 years ago and trying to get language models to produce consistently well reasoned results without fake information and it was near impossible. Today there are increasingly large datasets that are focused purely on reasoning and logic in specific domains, curated by human experts.

The models are fine tuned on these datasets to learn the patterns of reasoning and logic. So it can produce more reasonable answers and "explain" why it answers a question a certain way. It is still just pattern recognition but enhanced to mimic an expert instead of a lay person. It also learns to some extent what is garbage information and what isn't.

Once you get past the memes and trolls, Rokslide starts to look like one of those "expert" datasets. Here is what Artemis produces asking the question: Will a nightforce scope survive a 100 ft drop onto the concrete?

Based on the extensive testing and field experience described in the context, a Nightforce scope would likely not survive a 100-foot drop onto concrete.

While Nightforce scopes are consistently described as the most durable scopes available, with exceptional performance in drop tests, the context makes several important distinctions:

  • Standard drop testing is done from 18" to 36" (3 feet maximum) - this is considered sufficient to replicate real-world hunting scenarios
  • The author mentions successful drops from 7 feet onto rocks/asphalt/concrete with Nightforce scopes maintaining zero
  • Even extreme examples like a 60-foot drop off a cliff into broken rocks resulted in a functional scope that killed an elk the next day
  • However, drops above 6-7 feet on hard surfaces risk breaking turrets or eyepieces on any scope
The context specifically warns that "Done wrong, you will physically break any scope made" and recommends using soft ground with padded mats for testing. A 100-foot drop onto concrete would generate forces far beyond what any rifle scope is designed to withstand.

While Nightforce scopes have shown remarkable durability compared to other brands, a 100-foot concrete drop would likely result in:

  • Catastrophic physical damage to the scope body
  • Broken lenses
  • Sheared turrets or eyepiece
  • Complete mechanical failure
The answer is no - even the most durable Nightforce scope would not survive such an extreme impact.
 
Just came across this…mapping and text AI also pulling from Rokslide. All the more reason to stick to the forum rules. Interesting move from a Rokslide sponsor
I'm curious how everyone feels about their knowledge and experience being mined for profit... ?

I signed up so long ago I can't remember the "Terms and Conditions" of becoming a member of this site, but this seems to be another case of "if you are not paying for it the product is you".
 
I'm curious how everyone feels about their knowledge and experience being mined for profit... ?

I signed up so long ago I can't remember the "Terms and Conditions" of becoming a member of this site, but this seems to be another case of "if you are not paying for it the product is you".
I'll have to admit, I'm a little less than thrilled with my content being mined this way. Like you say, it's a free product though so I should know the risks.

What's hilarious is that someone may follow my sometimes misguided opinions because of it, lol..............
 
I'm curious how everyone feels about their knowledge and experience being mined for profit... ?
I see AI as us cutting our own throats. I'm opposed to it on multiple levels. But at the same time I fear the unseen implications of trying to restrict its use. So I am morally opposed to it, not legally opposed to it.

I can go back decades and show you where people were afraid of technology 'destroying hunting' and it really hasn't happened. The only real negative we've seen is that hunting has grown in popularity, and I don't blame tech for that. I blame the GWOT. That quickly and obviously goes far past the scope of this thread, of course.

From what I have seen, the best you can get from AI, is a sort of aggregator that can compile volumes of info and distill it down into something that sounds great but won't get you out of the truck and up the mountain and certainly won't put you on an elk.

In short, a massive amount of AI becomes mental masturbation for the people willing to pay for it. An entire summer spent online researching a hunting trip won't teach you a fraction of what you'll learn in two scouting days.

Also, I don't blame the people building and selling AI to the masses. They're just giving people what they ask for.
 
I was looking for replacement forks for a friend’s rotisserie and made the mistake of googling “spit roast” on a work computer.
I would not have expect that one to be a problem, guess I'm more innocent then I thought.
 
+1 @Chris in TN ... One of the big problems with AI is that diversity of thought (stick with me) is a real driver for learning. We'll never get to 'right' or 'best' answers unless we're willing to explore thoughts beyond the status quo (e.g. the recent prioritization of durability in scopes & bullets over headstamps). AI doesn't allow us to realistically entertain alternatives. Eventually the LLMs will be conditioned on content that is significantly AI generated resulting in a compounding feedback loop, death spiral, into absurdity.
 
I see AI as us cutting our own throats. I'm opposed to it on multiple levels. But at the same time I fear the unseen implications of trying to restrict its use. So I am morally opposed to it, not legally opposed to it.

I can go back decades and show you where people were afraid of technology 'destroying hunting' and it really hasn't happened. The only real negative we've seen is that hunting has grown in popularity, and I don't blame tech for that. I blame the GWOT. That quickly and obviously goes far past the scope of this thread, of course.

From what I have seen, the best you can get from AI, is a sort of aggregator that can compile volumes of info and distill it down into something that sounds great but won't get you out of the truck and up the mountain and certainly won't put you on an elk.

In short, a massive amount of AI becomes mental masturbation for the people willing to pay for it. An entire summer spent online researching a hunting trip won't teach you a fraction of what you'll learn in two scouting days.

Also, I don't blame the people building and selling AI to the masses. They're just giving people what they ask for.
I'm not talking about AI.

I'm talking about the harvesting of our knowledge for profit. By signing up for the platform I assumed I would be commercialized via Marketing and Ads. I did not think I would have my "expert opinion" (IP?) sold off by site owners

***This assumes that site ownership had to give the OK' for Base Maps to use that data, which I believe they would. Also, this assumes a transaction between site ownership and Base Maps for that data. I could obviously be wrong about 1 or both***
 
+1 @Chris in TN ... One of the big problems with AI is that diversity of thought (stick with me) is a real driver for learning. We'll never get to 'right' or 'best' answers unless we're willing to explore thoughts beyond the status quo (e.g. the recent prioritization of durability in scopes & bullets over headstamps). AI doesn't allow us to realistically entertain alternatives. Eventually the LLMs will be conditioned on content that is significantly AI generated resulting in a compounding feedback loop, death spiral, into absurdity.
Wholeheartedly agree with the first part of this because that is how many folks use it today. Ideally though it should be a way to explore alternatives and surface concepts you haven't considered. That depends on how you use it and how its used in a product. Actually new or novel ideas I think will always be in the domain of humans.

Also I agree that there is a chance of these things generate too much of their own content and the steady state becomes something absurd or just very plain.
 
I'm not talking about AI.

I'm talking about the harvesting of our knowledge for profit. By signing up for the platform I assumed I would be commercialized via Marketing and Ads. I did not think I would have my "expert opinion" (IP?) sold off by site owners

***This assumes that site ownership had to give the OK' for Base Maps to use that data, which I believe they would. Also, this assumes a transaction between site ownership and Base Maps for that data. I could obviously be wrong about 1 or both***

Very fair question and I'll go specific to Rokslide and then the general lay of the land since I think a lot of folks don't realize how much the landscape of the "free" internet is shifting under us.

Ryan and I have talked several times on this and outside of us being a paid advertiser, the main exchange is BaseMap driving traffic back to Rokslide (through links) and, hopefully, Rokslide can embed our search engine to provide better search functionality.

Now if you look at all the large AI companies, who use Rokslide data and every other website to train their models, there is largely no direct compensation and no permission given. Some really big internet publishers do have contracts but not many. The implicit tradeoff has always been eyeballs drive ad revenue. Google crawls the site surfaces relevant websites, they make money advertising on the search page and then they send traffic to you and you make money when the person lands on your site.

For the last 12 months, they provide AI overviews, fewer people click through because they got the answer (or think they got the answer). So Google still sends traffic but not as much because they are an answer engine now. Then you have other AI companies (OpenAI) who are not search engines but have trained on all this data and provide answers. They arguable send no incremental traffic and likely take away traffic that could have come from the search engines.

A couple court cases are making their way through the system on the legality of all this. Who knows where it lands but in the end, I'd guess site owners are gonna get the short end of the stick. AI companies will be able to argue their results are unique enough that they don't infringe on copyright. Site owners won't clamp down on crawling because you still need the search traffic. Some 2nd and 3rd order effects will happen: maybe more ads, maybe less "free" and more paid features, maybe independent sites sell to aggregate publishers because they don't make enough.
 
Very fair question and I'll go specific to Rokslide and then the general lay of the land since I think a lot of folks don't realize how much the landscape of the "free" internet is shifting under us.

Ryan and I have talked several times on this and outside of us being a paid advertiser, the main exchange is BaseMap driving traffic back to Rokslide (through links) and, hopefully, Rokslide can embed our search engine to provide better search functionality.

Now if you look at all the large AI companies, who use Rokslide data and every other website to train their models, there is largely no direct compensation and no permission given. Some really big internet publishers do have contracts but not many. The implicit tradeoff has always been eyeballs drive ad revenue. Google crawls the site surfaces relevant websites, they make money advertising on the search page and then they send traffic to you and you make money when the person lands on your site.

For the last 12 months, they provide AI overviews, fewer people click through because they got the answer (or think they got the answer). So Google still sends traffic but not as much because they are an answer engine now. Then you have other AI companies (OpenAI) who are not search engines but have trained on all this data and provide answers. They arguable send no incremental traffic and likely take away traffic that could have come from the search engines.

A couple court cases are making their way through the system on the legality of all this. Who knows where it lands but in the end, I'd guess site owners are gonna get the short end of the stick. AI companies will be able to argue their results are unique enough that they don't infringe on copyright. Site owners won't clamp down on crawling because you still need the search traffic. Some 2nd and 3rd order effects will happen: maybe more ads, maybe less "free" and more paid features, maybe independent sites sell to aggregate publishers because they don't make enough.
Appreciate the candidness.
 
I'm not talking about AI.

I'm talking about the harvesting of our knowledge for profit. By signing up for the platform I assumed I would be commercialized via Marketing and Ads. I did not think I would have my "expert opinion" (IP?) sold off by site owners

***This assumes that site ownership had to give the OK' for Base Maps to use that data, which I believe they would. Also, this assumes a transaction between site ownership and Base Maps for that data. I could obviously be wrong about 1 or both***
If RS content was locked behind a paywall this may be more controversial but the entire content of RS is available to anyone with a keyboard and an internet connection, all the AI is doing is searching more efficiently. I don't particularly care for AI either but like others have said, once you hit the trail that's when the real learning starts and no amount of AI is going to shortcut that.
 
Very fair question and I'll go specific to Rokslide and then the general lay of the land since I think a lot of folks don't realize how much the landscape of the "free" internet is shifting under us.

Ryan and I have talked several times on this and outside of us being a paid advertiser, the main exchange is BaseMap driving traffic back to Rokslide (through links) and, hopefully, Rokslide can embed our search engine to provide better search functionality.

Now if you look at all the large AI companies, who use Rokslide data and every other website to train their models, there is largely no direct compensation and no permission given. Some really big internet publishers do have contracts but not many. The implicit tradeoff has always been eyeballs drive ad revenue. Google crawls the site surfaces relevant websites, they make money advertising on the search page and then they send traffic to you and you make money when the person lands on your site.

For the last 12 months, they provide AI overviews, fewer people click through because they got the answer (or think they got the answer). So Google still sends traffic but not as much because they are an answer engine now. Then you have other AI companies (OpenAI) who are not search engines but have trained on all this data and provide answers. They arguable send no incremental traffic and likely take away traffic that could have come from the search engines.

A couple court cases are making their way through the system on the legality of all this. Who knows where it lands but in the end, I'd guess site owners are gonna get the short end of the stick. AI companies will be able to argue their results are unique enough that they don't infringe on copyright. Site owners won't clamp down on crawling because you still need the search traffic. Some 2nd and 3rd order effects will happen: maybe more ads, maybe less "free" and more paid features, maybe independent sites sell to aggregate publishers because they don't make enough.


To some extent this is simply the dynamics of any new frontier, with some degree of ungoverned spaces while the people and systems work out right, wrong, and legal. It's a gold rush. And, frankly, I'd rather have an American with a shooting, military, and hunting background staking this claim, than some Chinese or Russian profiteer. Especially one willing to engage here, with us, directly. This is a level of accessibility and even accountability I don't think many communities could ever expect. So, thank you.
 
Back
Top