The very first glimpse of AI run-out intelligence

Typical for most AI pimpers nowadays. They say it is epic, generational, incredible, never seen anything like it. Then you point out the obvious flaws and the response is it’s early, that’s a difficult problem, we’re working on it.
Well, it's true. Most people haven't seen something like this. They've seen projection systems, but not ones with ball tracking and scorekeeping. 100x more complex, made possible by AI.

As such, the model improves through more training data. Which takes time.

Welcome to software development.
 
Well, it's true. Most people haven't seen something like this. They've seen projection systems, but not ones with ball tracking and scorekeeping. 100x more complex, made possible by AI.

As such, the model improves through more training data. Which takes time.

Welcome to software development.
I'm intrigued by your progress on this new frontier! Hopefully you will have an API implementation soon so I don't have to build this myself. :)
 

I'm starting to do a "Drill of the Day" series, showcasing how easy it is to create your own drills and play them.

Another very cool feature is the drill's leaderboard. Drills have leaderboards, where you can submit a placement run. Your run through the drill is recorded and uploaded automatically, as well as the score and the time it took you to complete. This places you on a competitive leaderboard with other FusionCue users, where you can see exactly how they approached the drill.
 
Yup, still early. Doable, soon.

I have been working on a rack solver. I agree with Gary's sentiment that it is still early and takes time, but I have been putting in a lot of work and have some decent progress - but this is far from simple. At some point I'll release this app, give it a go, currently living at gotshape.net... click "Play runout"

10-ball solver runout
 
AI can't feel the humidity in the room, look at the table and see that the cloth is dirty and see that the table has a slight roll to it
 
I have been working on a rack solver. I agree with Gary's sentiment that it is still early and takes time, but I have been putting in a lot of work and have some decent progress - but this is far from simple. At some point I'll release this app, give it a go, currently living at gotshape.net... click "Play runout"

10-ball solver runout

i did. it played position on the 10 like no sane person would ever do.
 
Yeah, again, early days. Not sure this is "ai bullshit" as such, but it is software and algorithms, and I get the Luddite sentiment. People said similar things with chess in 97 with Kasparov v Deep Blue. I'm having fun for the last few weeks playing with exactly that problem, "dude, why would you try to get shape on the 10 in the side, that's loopy". This is AI one-shot "solve pool, go!", in my mind it is more encoding all the great knowledge that real pool players have, and it's interesting for me to think about the "why" of various shots and strategies. I think it plays well into learning/training.
 
When will the time arrive, if it hasn’t already, that a cue maker can utilize AI for cue design requirements, new creative designs and interface that with his other equipment like CNC. Imagine just asking for a design theme and seeing what AI comes up with and of course, it has to be attainable. So just seeing it displayed is worthless if you can’t build the cue design. Anyway, I am sure someone is tinkering with that right now, maybe start with a Bushka or Box Cue Design.

I dunno know where our world is headed for a number of different reasons and factions. However, as far as the pool community is concerned, I envision wondrous, amazing improvements never before imagined or contemplated await.

This stuff is pioneer engineering. The Wright Bros. were stubborn minded and JFK was inspirational, thank God. The
future can can be changed and shaped by anyone who has vision, talent & burning desire to push them to succeed.

It could be you, another reader, someone just born in Chicago or anyone for that matter if you are relentlessly pursue
your dreams. But it’s written and accepted that even the wisest of men knows their limits. So I know I don’t possess
that drive to push thru the hardest of challenges, adversity that cause you to question why go on. Just let it die. I had that experience trying to obtain a trademark 8 years ago. I just ran out of funds to keep pursuing it because a trademark attorney costs $400/hr. When there are other similar trademark requests preceding yours, you need a lawyer.
That's happened at least once.
 
Yeah, again, early days. Not sure this is "ai bullshit" as such, but it is software and algorithms, and I get the Luddite sentiment. People said similar things with chess in 97 with Kasparov v Deep Blue. I'm having fun for the last few weeks playing with exactly that problem, "dude, why would you try to get shape on the 10 in the side, that's loopy". This is AI one-shot "solve pool, go!", in my mind it is more encoding all the great knowledge that real pool players have, and it's interesting for me to think about the "why" of various shots and strategies. I think it plays well into learning/training.

encoding real player knowledge would have it watch hundreds of hours of video. this looks more like vibe coding tbh.
 
encoding real player knowledge would have it watch hundreds of hours of video. this looks more like vibe coding tbh.
Watching videos encodes player behavior, actions, and patterns, not player knowledge. It then creates the AI knowledge base, but that isn't the same as the player knowledge base.

This is an interesting endeavor and there will be a great deal more to come. It causes me to reflect on the research reports of AI doing surgery, which is spectacular, but it didn't actually do surgery.
 
I'm picking up on the pejorative language you're putting down. This has been a lot of fun work, and it is a hard problem. I like hard problems. If there is another all singing all dancing solution I've overlooked let me know, but "don't even try" is silly
 
The negative reactions are to the toxic blend of 1) carnival barking, over the top descriptions of what has been accomplished, 2) self-important inflation of the amount of work you put in, and 3) defensiveness whenever flaws are pointed out.

No, this is not the 'very first glimpse' of computers figuring out how to run out a pool table. No, it is not revolutionary to make a web app. It's just revolutionary for random joes who can't code to pay OpenAI $20 or $200 a month to have their frontier models do it for them.

Wait until your side project is ready for public consumption and then perhaps you will blow people's minds like you want.
 
The negative reactions are to the toxic blend of 1) carnival barking, over the top descriptions of what has been accomplished, 2) self-important inflation of the amount of work you put in, and 3) defensiveness whenever flaws are pointed out.

No, this is not the 'very first glimpse' of computers figuring out how to run out a pool table. No, it is not revolutionary to make a web app. It's just revolutionary for random joes who can't code to pay OpenAI $20 or $200 a month to have their frontier models do it for them.

Wait until your side project is ready for public consumption and then perhaps you will blow people's minds like you want.
Not sure if this was directed to me or ryan?

Either way, FusionCue has just about 100 paying users now and we're growing every day.

I will be sharing more news shortly, but we're all having super fun challenging each other with daily challenges in a 3D arena environment (which you can actually check out here: https://fusioncue.com/daily-rack)

I know some people have strong opinions who've never used it before. Don't ask me, talk with my users: https://discord.gg/AsXWJ73hBz

1794504248127.png
 
Last edited:
I'm picking up on the pejorative language you're putting down. This has been a lot of fun work, and it is a hard problem. I like hard problems. If there is another all singing all dancing solution I've overlooked let me know, but "don't even try" is silly

vibe coding isn't pejorative. it's also not impressive, as everyone and their mothers do it these days. i have barely literate friends proudly sending games/apps to me in whatsapp. it's rubbish
 
Pretty cool.

A couple thoughts/wishes...
- Show two (or more?) options at once so they can be visually compared
- Give reasoning for choices

pj
chgo
 
vibe coding isn't pejorative. it's also not impressive, as everyone and their mothers do it these days. i have barely literate friends proudly sending games/apps to me in whatsapp. it's rubbish
Vibe coding is a very loaded term. Not everyone who uses AI to develop things can be put into the same basket.

I've been a designer/developer for nearly 3 decades, and trained both designers and developers on a channel with 1 million+ subscribers and over 1,400 videos and nearly 100+ courses. Primarily UI/UX design and frontend/some backend development.

- There's your average joe who doesn't have a tech background, nor any experience, who can use AI to build software.
- Then, there are people who use AI to develop that have decades of previous experience without it.

These two are not the same, and will not produce the same results. As great as modern AI models are, their output still benefits massively by a trained eye in both UI/UX, and development.

And if you're a skilled developer who still prefers to do things by hand? Good luck, other skilled developers who know how to handle agentic systems will run laps around ya.
 
Your anxiousness is valid because we have never seen such a large change in technology during the lifetime of a generation.

Really? Electricity. Cars, then airplanes. Telegrams, radio, radar, television, cell phones. Space travel, then satellite communications. IBM, then personal computers. The internet.

This is just the latest thing, and it stands on the shoulders of centuries of technological advancement that changed the world at light speed. Watch the series 1923. See horses and cars side by side, people with and without running water and electricity.

If anything, technological change has slowed down in this century. Science needs to get with it. We need robots and antigravity, now.
 
The shape on the three drawing across the shot line was not good and your result proved it. Only missed by a few inches and the three became way too hard of a shot. Then a path rebounding off the side rail a ball width from the center pocket to get on the four is just stupid. Might be artificial but intelligent not much.

That needs to come a long ways to be competitive over time.
 
Back
Top