Joe Looking For Your Thoughts - The Challenges and Opportunities of AI (1 Viewer)

[Joe Bryant]

I asked in the FFA but also wanted to ask here.

Our GB @GroveDiesel recently posted:

Maybe I’m just an old man yelling at the clouds, but I mostly hate AI. I’m sure it has some great uses, but it’s sucking the humanity out of so much. I used to enjoy reading good sports writing and even game recaps that gave you the flavor of the game. Now you go read a game recap and it’s just clearly AI garbage spewing out a few stats and using analytics to point out a few key things in the most boring way possible.

I think lots of people feel like him. I know I do.

At Footballguys, I see AI as both a challenge and an opportunity.

The challenge of course is competitors now can create much more content at low to no cost using AI generated or AI assistance. Where we have a large staff of paid people, the opportunity to create by AI is here and many will take advantage of it. As GroveDiesel mentioned above.

The specific opportunity we see is directly related: In a world filled with more and more AI created content, we have an opportunity to stand out leaning into our staff and the community. At least that's my thought.

We're constantly asking, "What can we do that AI can't do?". That gap is of course narrowing, but it's still significant.

I'd love to hear from you folks on how you see the fantasy football space and how you see AI content in this space.

And specifically, what you think we could do to stand out.

Much thanks.

 

[SeniorVBDStudent]

I had a well thought out response to the philosophical question, and then I realized you were only talking about AI in the fantasy football space. I further realized this is a "no politics" arena, so I will refrain from personal bias influenced commentary on the larger question. To your specific question, I would say hire a consultant. That's a little flippant....perhaps others can give you more meaningful feedback.

 

[taylormeetstheismann]

AI will change the landscape. Considerably.

 

[osubuckeyeman]

There should always be a place for paid staff—personality matters. There are people I listen to because I like listening to them. I relate to that person through their podcast or, in this case, their writing. People have opinions, and whether you agree or disagree, it's beneficial to absorb the article and form your thoughts around that person's perspective.

If I chose to have a subscription to consume sports information, I would pay a little more for content written by real people. I do not mind a mix of AI, such as for statistics and metrics, but replacing people would not work for me.

I'm on this forum all the time because we are the writers—real people writing matters.

 

[Joe Bryant]

I had a well thought out response to the philosophical question, and then I realized you were only talking about AI in the fantasy football space. I further realized this is a "no politics" arena, so I will refrain from personal bias influenced commentary on the larger question. To your specific question, I would say hire a consultant. That's a little flippant....perhaps others can give you more meaningful feedback.

Thanks. I think this can be easily a non political comment. And I doubt we have the budget for hiring consultants. Would love to hear thoughts from the super smart folks here like you that are close to the topic.

 

[Ray_T]

as someone who has used AI to help write papers for my work, I have found that its ok in terms of getting that first draft done. it does an ok job that way but that first draft always requires editing.

If you want the product to be good you need a human to proof read it and add relevant information that an AI wouldnt find online. for your purposes, I'd say you still need someone who watches a good portion of the games and provides an opinion to give context to what the AI is spitting out.

AI can be helpful in terms of spouting analytics based stuff quickly and efficiently. but if you want to comment on something that requires the eye test, AI cant do it. so if someone misses a key block on a 3rd down play in a certain game, AI likely wont pick that up unless someone has commented on that in an article already. even then, maybe not.

but if writing a game summary an AI assisted first draft can be helpful to get you started and then you can add in personalized info from what you saw in that game

 

[Drunken Cowboy]

Good writers can do a lot that AI can't. AI still isn't really I. It can just rephrase content from elsewhere. It can't offer insight.

 

[jasvic]

Limiting to news and information, AI is not a useful tool for writing and creating content, IMO. It isn't intelligent enough to know what is true, instead it parrots what it "knows." Do any Google search and see the AI result at the top. It's not uncommon for that info to be incorrect.

As an example, take a polarizing player like Jalen Hurts. Many still think he's a bum while others say he's a top 10 or top 5 QB. What would AI say about it? It can't offer insight or opinion, it will crawl the opinions of thousands of people/stories and come up with some amalgam of it.

And of course you lose out on personalities and styles of human writers.

I'm sure it would have uses when dealing with absolutes, like statistics, but any news organization relying on AI to write any content is not one I want to read.

 

[Bri]

I've written spider/crawler scripts and followup scripts to analyze the data for decades.
I haven't had my site up for about 10 years. I still code. I still tweak things.
AI is powerful and simpler. Like replacing a gigantic 25 MB memory card of 1990 with a teenie tiny 1gb card of today.
I don't have to write much code anymore.

I don't share my results. I use them for me and me alone.
I think AI is going to eventually be like DOGE and some of my friends are going to lose jobs or respect.

What about the wonderful writer whose rankings seem great but at the end of the year the results show they were quite poor? What about the draft guy I love to read but his prospect ranking are way off?

100 years ago, Drinen and Chase wrote for PFR and masterfully proved or disproved theories.
In a similar but lesser fashion, we used to write about theories and then cherry pick results that prove it. RSFF or OL yeller would have people saying "wait a sec you're cherry picking!" And we'd debate theories further and...a lot of that died. I'm still a nerd doing that.

I've failed mightily to use scripts and AI to help me figure out dfs. I've stopped for a couple years now. Someday I'll get back to that.

I invented the mock draft simulator and David Dodds provided incredible feedback. What if this happens? Yeah but you didn't incorporate this? So I thought of some and went back to the drawing board and came out with the script. Now everyone uses my script or a variation and that IS how the Internet works. I understand.
The draft dominator was glorious. Absolutely loved it and it was clearly better and more robust and all sorts of glorious things. I've had lots of conversations about theories and stuff about this too.
This is AI that's been in FF for decades. This is our example.

I can wreck just about any sites model and draft awesome against it if I get enough tries. There's a few but mostly I don't think they know how or why I did this n that or FBGs did this or that. I am not telling them either.

I sincerely believe I am the best drafter and can win any league and have never met anyone with near as good a record as I.

One of the things I can do well is predict who you'll take next.
Why?
Because I follow you and read your articles and listened to your show and the guy you mentioned 30 times will reasonably go at your pick.

There's more.
If I played with you in a few leagues, I have notes on your drafts and trades.

Draft rating tools also tell me who you like if I plug enough drafts in and skew the rankings enough times.

A few years back, my buddy was in the hospital and he said draft who you think I will. His wife was to give me a paper he wrote on but she was upset and forgot. Welp he made a list of 30 guys and I drafted 19 of them in a 20 round draft. He thanked me for using his list and the wife asked me not to say. All good, bigger problems to deal with etc.

I can go on and on with weekly waiver wire stuff.

Auctions are so easy to predict when you've been with people for years that I stopped tracking it.

So I get invited into a league and say I will win it. I'm not being cocky. I was a FSWA HOF finalist the first few years and this is what I'm good at. Usually there's a day or two delay and then they either want me or don't. They all want to beat the bragger and I literally say almost nothing and just take the prize money.

There's gut calls. AI will never beat this.

There's human nature in this. I know people that feed the masses with rankings and don't necessarily believe what's put out. They'll get exposed when their boss or followers see they never draft the guys they raved about.

AI is going to force people to be brutally honest. Just say ya messed up or take your flowers. I don't know that our world is built for this. Many are guarded of their flaws and think it'll hurt their reputation.

The AI articles are easy. It's just mad libs like. The app game changer does it for high school sports. I wrote some awful articles in my day and got beat up pretty good for em. I learned techniques and styles and eventually wrote for USAToday. From my POV, if you're going to write you're going to have to learn the same stuff because madlibs articles are going to be so plentiful.

For stats-
I can highlight a chart and plug it into AI and have it write an article too. So don't do that. Don't write sentences that just regurgitate the chart. Tell me how you feel. Tell me how you'll use this info. Tell me how my FF team will be better for me reading this article.

I know this sounds braggy. I have made about 100,000 mistakes in FF since I started playing in 1990. Do you know how many had waivers and bad trades and bad evaluations I've done? I've learned from them and some were hard to do.
I believe the best players, coaches, teams, FF people, and writers all made mistakes and learned from them. I adore these people so much. It's how coaches and parents and teachers taught me and I believe it's the way.

Let me give you an example. I drafted Kurt Warner in the 20th round the year he broke out. We say that right? He was my third quarterback that draft. I didn't even want three quarterbacks but my backup had an injury and I was OMG fine give me that guy. I didn't draft Kurt Warner expecting greatness as it often gets explained.

My first pick ever? Dan Marino. Because I was 16 and had a couple beers and that's what came out when I spoke. After I used a note paper I had in my pocket and put thought into it and stopped drinking but woohoo I drafted a gem and give me credit for that.
I could write so much more of my mistakes.

I think Joe has the most integrity of any man in FF and possibly that I've ever known of. I have absolutely zero doubt that FBG has any hiccups with AI being used to track things or how it's handled or anything. There isn't another Joe in the world though.
Scott Fish- an absolute gem of a man. Zero doubt.
I can go on and on...it's going to be interesting watching humans be humans.

I think a lot of sites need Joe leading them and will wish they had him. I think y'all that frequent this board for a couple decades like me know exactly what I mean.

The people that do get exposed for not doing well. Guess what? Do well tomorrow. Get another job and adjust your thinking and rock at the new job.

The industry is going to be tweaked hard but it doesn't have to be doom n gloom. It can be a means to make everything and everyone better

 

[IHEARTFF]

Did AI write that

 

[jwb]

it did write this in 10 seconds:

me: can you tell me why footballguys is a premier fantasty football site and why using human writers like they do is an advantage?

ChatGPT said:​

Absolutely—Footballguys is widely considered one of the premier fantasy football sites because of its depth, consistency, and human-driven insight that goes far beyond algorithmic projections. Here's why it stands out:

---

​

Footballguys has been a trusted name in fantasy football for over two decades. It’s not just the massive database of rankings, cheat sheets, and projections—they deliver a full ecosystem of content: matchup analysis, dynasty rankings, IDP content, DFS breakdowns, customizable tools, and perhaps most importantly, context-driven advice that adapts week to week. Whether you're in a 10-team redraft or a 32-team dynasty superflex, Footballguys has tailored content for you.

They’ve also built a loyal community that relies on insightful, long-form analysis, not just clickbait or AI-generated blurbs. Their Draft Dominator app is elite. Their newsletters are timely. Their forums are stacked with hardcore fantasy minds. It’s a site made by fantasy nerds for fantasy nerds—and it shows.

---

​

While AI-generated content is fast and data-rich, human writers bring nuance, instinct, and experience to the table—especially in a game that’s equal parts strategy and psychology. Footballguys’ writers can see when a player’s usage is trending up, even if the raw stats don’t yet reflect it. They watch film. They follow beat reporters. They understand coach speak. They know how to read between the lines when a team says “RB by committee” but really means “this guy is our closer.”

Human writers also help synthesize conflicting data points. Algorithms may not know that a wide receiver is playing through a nagging hamstring, or that a cornerback he's facing has quietly become elite. Humans do. And Footballguys employs some of the best in the game—from Joe Bryant to Sigmund Bloom to Matt Waldman—who consistently give actionable advice with sharp, experience-based angles.

 

[Donnybrook]

Grok AI mock draft had 4 QBs gone in the first 9 picks. My thought on this is AI is factoring too much into top teams need, measurables and player stats. AI might not be able to process the on field performance of players.

One of favourite writer was talking about Malachi Starks recently. He said that is on field speed is better than his combine/timed speed and measurables. Compared Starks to Brian Branch in that respect. Most scouts raved about Branches on field performance and football IQ. They knew that his average combine numbers didn't capture him accurately as a player.

 

[Jayded]

AI is best used for mundane tasks overall. Avoiding it as professionals will be to your own detriment, but those effective at using it will otherwise pass you by.

Have it do the math or summarization of free text or scanning sentiment of many data points….but then involve the human touch post results. AI isn’t one of us that enjoys the hobby and that sentiment is still a ways off.

 

[kutta]

I’m in a similar conundrum with my business. We write a lot of proposals for work. Our competitors are starting to use AI to write their proposals, and honestly, they are pretty darn good.

Long story short, we’ve realized that in order to compete we are going to have to use it. We’ve taken the approach of using it as a starting point, but making sure we go through and change and tweak and add everywhere we think we need it.

I think it’s a mistake to think the “human touch” will win in the end. Competitors will use it and get material out much quicker than you can even imagine. Right now, it seems like a combo approach is best - add your personal touch to what AI can get started for you.

 

[JAA]

I’m in AI for my day job. You have nothing to worry about.

AI is miles wide and 1” deep. GenAI is currently at the novel phase, yes helpful, but not contextually aware.

Conversations with opinions and biases and feelings are what will keep people coming back here. It’s not just content, it’s context and people.

 

[BassNBrew]

I’m in a similar conundrum with my business. We write a lot of proposals for work. Our competitors are starting to use AI to write their proposals, and honestly, they are pretty darn good.

Long story short, we’ve realized that in order to compete we are going to have to use it. We’ve taken the approach of using it as a starting point, but making sure we go through and change and tweak and add everywhere we think we need it.

I think it’s a mistake to think the “human touch” will win in the end. Competitors will use it and get material out much quicker than you can even imagine. Right now, it seems like a combo approach is best - add your personal touch to what AI can get started for you.

Posted this in the FFA, but AI is touting a Josh Gordon comeback right now.

 

[BassNBrew]

i heard he's (Gordon) coming back for Cleveland in 2025

Yes, the buzz is true—Josh Gordon is making a comeback with the Cleveland Browns. According to NFL Network's Tom Pelissero, Gordon is returning to the team and will gradually ease into football activities. While there's no set timetable for his return to the field, the Browns are taking a cautious approach to ensure he's fully prepared before participating in games.

 

[Ruffrodys05]

i heard he's (Gordon) coming back for Cleveland in 2025

Yes, the buzz is true—Josh Gordon is making a comeback with the Cleveland Browns. According to NFL Network's Tom Pelissero, Gordon is returning to the team and will gradually ease into football activities. While there's no set timetable for his return to the field, the Browns are taking a cautious approach to ensure he's fully prepared before participating in games.

Ummm......okay. This tells me all I need to know about AI at the present time.

I do not know if it has been all of the "artificial intelligence goes drastically wrong for humans" type movies I have watched over the course of my life but, I have a very healthy fear of any type of AI being created and taking hold in our world. I believe it is a huge mistake to create artificial intelligence. At some point in the near future it will evolve past the point of being controlled by humans. Nope. Not. Interested.

On the other hand I fully understand the appeal of AI. Humans are always searching for ways to make their lives easier or more productive. Any assists on that front have always been welcomed by humanity. I do not see any difference with AI as I think most of humanity welcomes their eventual intelligencia Overlords to the world. Unavoidable at this point. I hope I am long gone before those Overlords take control of the planet.

 

[CalBear]

All AI can do is generate template content. For some writing, that's enough, but for an analysis site it's kind of useless. It can't do analysis; it can only tell you what other analysts are saying. So if you want to generate lots of plagiarized content about what's going on in the NFL, fine, it'll do that, but I don't see how that's interesting or in keeping with what FBG has been.

 

[Ilov80s]

Here’s one of the challenges, META AI is using its licensing agreements with celebs such as John Cena to engage children in sexually explicit chats. Fooling the kids into thinking that John Cena wants to meet them for illegal sex acts.

Justin Ling (@justinling.ca)

Good god. Meta is using their AI licensing deals with John Cena, Disney, and others to peddle sex chat bots to kids. Burn this company to the ****ing ground. https://www.wsj.com/tech/ai/meta-ai-chatbots-sex-a25311bf

bsky.app

 

[Gatorade]

Here’s one of the challenges, META AI is using its licensing agreements with celebs such as John Cena to engage children in sexually explicit chats. Fooling the kids into thinking that John Cena wants to meet them for illegal sex acts.

Justin Ling (@justinling.ca)

Good god. Meta is using their AI licensing deals with John Cena, Disney, and others to peddle sex chat bots to kids. Burn this company to the ****ing ground. https://www.wsj.com/tech/ai/meta-ai-chatbots-sex-a25311bf

bsky.app

WTH ?

 

[GregR]

My experiences are similar to what others have said. It is often good to get me started on something, and I can take what it spits out and start editing and revising and get to a result I'm content with in less time.

Some of the obvious things like stat gathering, it could do a good job. When I was writing for FBGs, the first step of my kick and punt returner predictions was farming a lot of data off the web. Vegas point spreads, kicker touchback rates, average number of team drives per game for the season and more recently, etc. Dumped into a spreadsheet and run through a lot of formulas, they would spit out a first pass picture of the game. How many scores, how many drives likely to end in a punt or on downs or result in a kickoff. I would then go through in more detail and adjust for other factors that I didn't think were well represented in the source data, or where I just plain didn't agree with Vegas on a game for example.

I could see having used AI to help me pull all of that together in a fraction of the time it took, so I could spend more time reviewing articles, watching video of returns, etc. I could also see using it as a research tool. It was very difficult sometimes to find comments about returners since they aren't usually much of a focus. Being able to add an extra layer of search capabilities there would have helped tremendously. But it wouldn't have replaced drawing conclusions as player situations changed, it wouldn't have replaced watching returns and noting players whose stats didn't show it but who frequently looked dangerous but got tripped up just short of breaking big plays.

Add me to the voices of the others saying, don't ignore it. Make use of it, but differentiate yourself by ALSO bringing in those human elements that it is very poor or unable to duplicate.

 

[davidjhonda]

Didn't see this thread earlier. Sorry for the late response.

I haven't used AI in fantasy football yet. Maybe I should and I wouldn't LOSE ALL THE DAMN TIME. But the comments about pulling together stats quickly are 100% correct.

But I use it extensively in my business. And, it's been a life-changing experience.

In my opinion, AI does two things very well:

1. It gives us knowledge. We could Google for hours and hours to get the knowledge we need. ChatGPT provides it in seconds after you know how to ask for it.
2. It leverages our time. I think everyone understands this. Once you know what you want, AI quickly cranks out the stuff.

Here's a description of how we used AI effectively. We wanted to boost our Google Organic rankings for various search terms. ChatGPT helped us understand exactly what we needed to do. We fed it information, and it made the right suggestions. One of the key suggestions was to do a blog highlighting specific keywords. ChatGPT wrote the content in seconds. We had to tweak it a bit. But, after we tweaked the first article, it remembered and applied the tweak from that point forward. We created 50 articles in 1 day. Before AI, that's easily a three-month project.

We could have done all this ourselves, but we don't have the time.

This brings me to my next point. We are the right size to implement AI. There are 18 of us, and we're big and sophisticated enough to know what we want. But we're always limited by a lack of knowledge and time.

Large companies have more resources, so they often already have the right knowledge and manpower. Very small companies often don't have the sophistication to know what they want.

So, whether it's right for fantasy football, I don't know. But it's perfect for businesses of a certain size.

 

[massraider]

So, whether it's right for fantasy football, I don't know. But it's perfect for businesses of a certain size.

Agreed.

My fantasy content consumption has moved beyond 'one guys opinion' to opinions backed by analytics. The guys who do that hold my attention.

And the work that goes into analytics, I respect. I have to think AI makes that much easier. Someone who has data backing up a stance is worth 100 guys on Twitter posting, "I don't know who needs to hear this, but..."

 

[Hot Sauce Guy]

I think lots of people feel like him. I know I do.

Add me to the list.

That said, I’m more like an artist I saw on Twitter who said, “I want AI to do the dishes & fold the laundry so I can be freed up to write my novel & paint, not an AI to write my novel & paint while I do my dishes & fold my laundry.”

 

[Chaka]

Im becoming a big fan of chatGPT, I think the biggest stumbling block for many people is understanding that the quality of the information AI provides is almost entirely dependent on the quality of information the user provides in their query.

Garbage in; garbage out.

I am enjoying using chatGPT to the point that, I am going to float the idea to my main fantasy league, which is 10 teams, that we add at least one AI managed team. Personally I would add two, one in each division.

As many have stated, you still need humans to manage the output of AI, but if you take time to curate it, you can get amazing results. And the time you put in is well pales in comparison to the value you can extract from it.

 

[Chaka]

I may use chatGPT to help craft my subscriber contest team.

And, I doubt I'll be the only one.

 

[Hot Sauce Guy]

AI provides is almost entirely dependent on the quality of information the user provides in their query.

Garbage in; garbage out.

This is a bit of a misnomer.

GIGO is a classic database acronym. Garbage in, garbage out. That part is correct.

With ChatGPT, it’s not only the query that is the “garbage in”, it’s more the quality of data that ChatGPT is accessing to provide the response.

I’ve seen quite a few erroneous responses from ChatGPT, such as the somewhat infamous answer to the question “how many rocks should a person eat per day?” - the onion once wrote and article with this as satire, and AI picks up on that satirical article & provides an answer based in part on that.

That is just one example. There are good applications for AI - I’m not going to poo poo the entire industry. But it’s not seasoned/sophisticated enough to trust with medical, or even fantasy football advice at this point.

Maybe it can fill out my shopping list based on what my smart fridge tells it I’m low on, or based on a meal I tell it I want to prepare.

But on the whole I believe people are putting way too much faith into it very prematurely. And there are a ton of folks out there deliberately throwing up junk to confound AI scrapers because the companies behind them are unscrupulous and don’t respect trademark or copyright protection.

And on that front, a FF advice response to a prompt might well contain scraped opinion from multiple FBG contributors without attribution. That’s probably my biggest issue with AI.

 

[Hot Sauce Guy]

I may use chatGPT to help craft my subscriber contest team.

Probably a fine use of AI. I’m not sure you’ll get quality results, but then you might have had worse results based on your intuition/football knowledge.


That said, I’d be a little concerned that AI will see it as more of a math problem in doing the budget than a strategic challenge, which is more the nature of the contest.

I’d be interested to see your team if you do that.

 

[Scoresman]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

 

[Hot Sauce Guy]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

 

[Chaka]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

That really depends on how I craft my prompts, doesn't it?

And if I leave it alone after a single prompt or, as everyone should when using these tools, if I refine the prompts as the answers come in.

 

[Hot Sauce Guy]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

That really depends on how I craft my prompts, doesn't it?

And if I leave it alone after a single prompt or, as everyone should when using these tools, if I refine the prompts as the answers come in.

Possibly. It also depends on what the AI is scraping to get your answers, but yes - if you’re incredibly specific about your inputs then sure.

That said, if you’re that specific with your inputs, you don’t really need AI.

 

[joey]

Based on my experience so far with getting “accurate results” out of AI like Gemini, I’d guess that if you use AI to try to craft a contest team, it will not be able to total the salaries to keep you under $200. It’ll say it did, but it won’t.

AI is a fantastic eager intern without actual experience and knowledge that YOU have. So you take it’s results and then spend time correcting its errors, or it’s emphasis on trivial points you didn’t want to make, but it definitely saves time in the right circumstances.

 

[Chaka]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

That really depends on how I craft my prompts, doesn't it?

And if I leave it alone after a single prompt or, as everyone should when using these tools, if I refine the prompts as the answers come in.

Possibly. It also depends on what the AI is scraping to get your answers, but yes - if you’re incredibly specific about your inputs then sure.

That said, if you’re that specific with your inputs, you don’t really need AI.

I think you are underestimating the power of these tools.

They absolutely can hallucinate, but to dismiss them as mere "scrapers" doesn't acknowledge the power of these tools.

And "tools" is the appropriate term. You can use a hammer to help create amazing things, but if you treat every problem as a nail you won't get very far with your hammer.

Like any tool you need to learn how to use it properly. But, if you do it will absolutely and unquestionably make your tasks easier and your outcomes far more reliable.

 

[Hot Sauce Guy]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

That really depends on how I craft my prompts, doesn't it?

And if I leave it alone after a single prompt or, as everyone should when using these tools, if I refine the prompts as the answers come in.

Possibly. It also depends on what the AI is scraping to get your answers, but yes - if you’re incredibly specific about your inputs then sure.

That said, if you’re that specific with your inputs, you don’t really need AI.

I think you are underestimating the power of these tools.

They absolutely can hallucinate, but to dismiss them as mere "scrapers" doesn't acknowledge the power of these tools.

And "tools" is the appropriate term. You can use a hammer to help create amazing things, but if you treat every problem as a nail you won't get very far with your hammer.

Like any tool you need to learn how to use it properly. But, if you do it will absolutely and unquestionably make your tasks easier and your outcomes far more reliable.

You said it yourself. GIGO.

They’re not just scrapers. I never said that. But the data they utilize has been scraped. And they get the good with the bad.

I’m somewhat familiar with the mechanisms & capabilities of AI. I’m not convinced it’s ready yet. It’s currently being oversold by folks who have vested interest in seeing it be widely adopted.

I’m not totally anti-AI. I just don’t believe it’s a fully baked tool.

 

[Chaka]

ChatGPT is going to pick up all the "20 Kicker team" jokes made here over the years and tailor your optimal team accordingly. It'll probably Bump Finley too.

Exactly.

That really depends on how I craft my prompts, doesn't it?

And if I leave it alone after a single prompt or, as everyone should when using these tools, if I refine the prompts as the answers come in.

Possibly. It also depends on what the AI is scraping to get your answers, but yes - if you’re incredibly specific about your inputs then sure.

That said, if you’re that specific with your inputs, you don’t really need AI.

I think you are underestimating the power of these tools.

They absolutely can hallucinate, but to dismiss them as mere "scrapers" doesn't acknowledge the power of these tools.

And "tools" is the appropriate term. You can use a hammer to help create amazing things, but if you treat every problem as a nail you won't get very far with your hammer.

Like any tool you need to learn how to use it properly. But, if you do it will absolutely and unquestionably make your tasks easier and your outcomes far more reliable.

You said it yourself. GIGO.

They’re not just scrapers. I never said that. But the data they utilize has been scraped. And they get the good with the bad.

I’m somewhat familiar with the mechanisms & capabilities of AI. I’m not convinced it’s ready yet. It’s currently being oversold by folks who have vested interest in seeing it be widely adopted.

I’m not totally anti-AI. I just don’t believe it’s a fully baked tool.

I'm not sure where we are in this conversation, probably talking past each other a bit. But no one is selling AI as some kind of autonomous substitute for critical thinking.

But, it's a very powerful tool and I think you're underestimating it's ability to weight the data it finds.

Again, it's more about the users ability to properly deploy the tool than anything else.

 

[davidjhonda]

Great comments.

Artificial Intelligence is an interesting term. To me, it’s much more a tool that helps us do our work much better and much faster. In my opinion, many people listen to the hype and have unrealistic expectations.

AI is all about the prompts. We can’t just paste a spreadsheet and ask it a question. When we use ChatGPT to do business analytics, we have to be very specific and say exactly what the data represents for each “column.” We also have to guard against erroneous results by saying things like “ignore this data if this condition exists.”.

It does scrape data but it adds much more. As a simple example, I was putting together a complicated Google Sheet with some very advanced formulas. I couldn’t figure out this crazy lookup from another workbook. I asked ChatGPT to write the formula and gave it the cell references. It spit out the perfect result. I used to read the Google help docs and solve this sort of thing via trial and error. But now I do it in seconds.

If I was in the fantasy business, I’d definitely use it to create content. But the instructions wouldn’t be “write an article on the Dallas Cowboys RB situation.” Then, I’d be basing the article on scraped content which I may or may not agree with.

Instead, I’d tell it the writing style like “positive and friendly so that a 4th grader could understand it.” Then, I’d tell it the length like 200-250 words and give it some guidance on the structure (intro, conclusion, etc.). And I’d tell it the points I want to make and the conclusion.

Chat GPT would write the article in seconds. And the writing is top-notch. Essentially, it reduces the 5 hour writing session to 10 minutes. Then, I’d have more time to brainstorm and produce more content.

AI is a tool that will change the fantasy business. Not because it “thinks” for us. But because it allows the good fantasy people to leverage their time and create 1000% more good stuff. At some point, those that embrace AI will win in this space. It’s inevitable.

 

[32 Counter Pass]

Reading the comments in this thread makes me think back to my life as a lithographer. In its heyday the industry employed a significant number of people in the Bay Area, and had a very strong, robust union. Then when desktop publishing started to be introduced into the industry I heard many of the same comments posted here about AI. People said desktop publishing could only do simple tasks. Or desktop publishing will never replace the etching side of the industry. Or desktop publishing will always need film to make plates. Guess what? The technology evolved and improved, wiping out the entire industry.

More recently, I recently retired (less than two years ago). In that time dramatic advances were made in my field to the point that most of the software that were industry standards has been replaced. Tasks that used to take weeks can now be done in minutes with these new tools.

If history is our guide, the AI technology out now is not the final iteration. It will improve and negatively impact every industry. AI is also going to be very, very disruptive in the FF landscape. The people that learn how to prompt AI properly are going to reap the benefits.

 

[zed2283]

I just started using AI a bit this year, mainly ChatGPT. I think that it's a great tool, and I think of it as a very advanced Google search. I like how it works and it's functionality.

That said, after using it for a little while the limitations become apparent. I tried to get it to give me some insights on my dynasty team, and it couldn't seem to stop giving me information that was basically two years old. I've seen similar instances in other queries where it gives me info that is obviously incorrect. So I take it for what it's worth and make my own conclusions.

Similarly, Watson has been used for how long on ESPN fantasy? And it doesn't seem particularly insightful.

So I think to Joe's question, I like AI when I'm searching for factual information because of it's speed and ease of use, but I don't want it trying to give me any color commentary, so to speak. We still need humans to contextualize.

 

[Harry Frogfish]

You can take it or leave it, but the gripes seem to be that it's not all there yet or has holes here and there? That's shortsighted and irrelevant, the simple answer is it will be. And this thing's coming faster than expected. You can adapt or get left behind, like it or not.

Horse hasn't just left the barn, he's long gone.

 

[32 Counter Pass]

All AI can do is generate template content. For some writing, that's enough, but for an analysis site it's kind of useless. It can't do analysis; it can only tell you what other analysts are saying. So if you want to generate lots of plagiarized content about what's going on in the NFL, fine, it'll do that, but I don't see how that's interesting or in keeping with what FBG has been.

I do not think this is true. A quick google search returns many examples of AI analysis. “Yes, AI can definitely assist with statistical analysis. In fact, many AI tools are designed specifically for this purpose, and AI is being increasingly used in statistical analysis for tasks like data cleaning, feature selection, and model building. ”

 

[jwb]

All AI can do is generate template content. For some writing, that's enough, but for an analysis site it's kind of useless. It can't do analysis; it can only tell you what other analysts are saying. So if you want to generate lots of plagiarized content about what's going on in the NFL, fine, it'll do that, but I don't see how that's interesting or in keeping with what FBG has been.

I do not think this is true. A quick google search returns many examples of AI analysis. “Yes, AI can definitely assist with statistical analysis. In fact, many AI tools are designed specifically for this purpose, and AI is being increasingly used in statistical analysis for tasks like data cleaning, feature selection, and model building. ”

You are correct - AI has gone far beyond simply repeating what was already said online.

 

[Ramblin Wreck]

Can someone explain to me like I'm 7 some cool things you've done with AI and fantasy football? I've dabbled around with some of this stuff but admittedly do not know it's full powers.

 

[32 Counter Pass]

I have dabbled as well, and am not expert, but my understanding is that the key to harnessing the power of AI is to know how to prompt the system correctly. If you simply ask what are the three most predictive stats for repeating as a top 12 RB, you will generally just get scraped info. But if you ask more detailed questions you get better results. It is relatively easy to find info about how to prompt AI correctly for better results.

Circling back to Joe’s post I think that sites that learn how to use AI correctly will uncover information that would be very time intensive for a person to do.