I asked ChatGPT about the influence of fuselage length and it returned the following table - which I hadn't seen before but seems accurate to me.

Could anyone confirm if the first column - regarding pumping and lift to foil - is also correct?


Thanks!

I was always told longer is better for pump and have always experienced that shorter was better for me (I use small tails). I think a lot of that is nuanced by the foil size, tail size, rider skill. The lift-off category is the only one that is a little questionable IMO. Perhaps there is something there because of the pumping but the green, yellow, red might not accurately describe how quick the flight is.

As small trick for ChatCPT:Its first answer is optimized for speed, saving energy (computational CPU) at the expense of accuracy.Just replying "are you sure" will make it activate its more powerful (but slower) mode, to get a better answer.

I have challenged its answers. It responds, you are correct, then goes into way more detail and corrects itself.

I love using it. I think it's better than the others.

Usually incorrect information from AI about foiling.

Select to expand quote

Retina said..
Usually incorrect information from AI about foiling.

Explain why

Select to expand quote

FranP said..

Retina said..
Usually incorrect information from AI about foiling.

Explain why

It's a new sport, so there's relatively little content that the AI was trained on. And what is there may be outdated (knowledge cut-off is a problem too for a new sport), or it may be forum posts where the writers don't give the correct answer, or the answer is a matter of preference.

Hey Fran table is good in line with my findings.
Would suggest including a 5th column- "upwind ability"

Copilot disagrees

Select to expand quote

Retina said..
Usually incorrect information from AI about foiling.

Plot Twist: Actual humans give also incorrect information...

I uploaded a video to it and it analysed my riding, check it out on YouTube @Dan everitt

I asked Microsoft Copilot why I was so crap at pumping a foil and it responded 'You're not "crap"-you're learning a highly technical skill. Pump foiling is notoriously difficult to master, and every faceplant is a step closer to that euphoric glide. Even seasoned foilers spend months dialing it in.'
It then went on to offer chatty advice summarised from a zillion pump blog posts, gave me advice on foil set-up for my KT Atlas and Nomad gear, shimming etc and gave me a list of drills to practice at the next session.
Hells bells!

The key to using AI optimally is writing as much detail as possible into your prompts. I recently used it (out of curiosity) to think through a complex navigation issue with downwind foiling, where I already have experience, and the response was remarkably good.

Try going into more detail with rider skill, foil size, board length, wing size, typical conditions, etc and see what you get.

Simple prompts tend to generate simple outputs.

Like others said, also get more granular with additional follow up prompts. I've definitely caught it hallucinating many times, but less now than 2 years ago.

Select to expand quote

hilly said..
Copilot disagrees

Regarding fuse length, I bet if you updated the prompt to include rider weight and size comparison of the fuses, it would be a different outcome. As a light rider, a 70CM fuse is much harder to pump from my experience in comparison to 60 and 50 cms. It just takes more lateral weight shift to get the same result as a heavier rider, who gets more results for similar movements.