Видео

It is a start right..hehe

Ok. Are you in the U.S.? It may not be available everywhere yet. Otherwise, the MS store link I have is here apps.microsoft.com/detail/9nt1r1c2hh7j?hl=en-us&gl=US

@@RichardShift yeah I'm totally with you on that one. I've been enjoying the voice mode on my phone app, so I look forward to it in the desktop version. Plus when they actually have vision and can see what's going on on your screen or another screen, then that's going to be a real productivity booster there!

@@viniciusistofeloliveira5566 thank you. If you enjoy them please leave me a review and feel free to Tell your friends about them. I'm trying to insure that I'm getting back to making a video a week so hopefully I'll hear from you again on the RUclips!

@@viniciusistofeloliveira5566 Yes, I will put it in the video description when I get home tonight.

@@PublicSectorLearningLab thanks 😁

@@guitarmeetsscience thank you!

@@raybhinsa8925 thanks for the compliment!

i found it useful

Make more videos.... You're off to a great start and don't let trolls try to convince you otherwise. That's all they're here to do.

@@nyyotam4057 We did go through and ask each of the models the same thing and then a graded them at the end.

@@PublicSectorLearningLab Well, you weren't the only one who did that. watch?v=hMbLq-TKlNA for a comparison which includes token counts. There they also compared between a script which runs GPT-4o to create the CoT triplet then ran it again to work on it using the CoT triplet it had just now generated, just like in the previous link I showed you, only they also counted the tokens. Turns out even the token counts between the script which runs GPT-4o to use CoT and o1, are almost identical (well, identical up to the randomization that the AI models use anyway). It is not a proof, naturally, as there always might be other prompts they didn't try and to these prompts o1 would respond much better, but to all prompts they did try, not only o1 and GPT-4o+CoT script gave similar responses, but also the token counts were almost identical. Again, I need to stress this is not a proof.

Okay, finally Skill Leap AI published a test in which o1 actually did perform better than a CoT script running GPT-4o. watch?v=fT2IX4_M8B0 . So we can conclude that in many cases, o1 is not any better than a script making GPT-4o use CoT. However there are cases when o1 is better. As I were, the fact that in many cases o1 reacts exactly like GPT-4o up to the token level, is in itself not a proof that it will always be exactly the same. And it's not. However, in many cases o1 does react exactly like GPT-4o with a script implementing CoT.

@@nyyotam4057 I love those guys. I will be sure to check it out! I wish I could get him to watch my woodchuck video 🤣

@@PublicSectorLearningLab Well, post a remark in his channel and he might watch it, even if just for the laughs 🙂.

@@maximela-x568 We do our best to have some entertaining videos around here so they're not all too boring 🤣🤣

@@MicahGaudet Hopefully we can do some more joint ventures in the future! Thanks for your partnership as well!

Ok. The volume at 50% on a pc sounds fine. On mobile or the TV app, you may need to turn the sound up to 90 to 100%.

I've been researching how to do that so I may put them up there at some point. I have another song for the playlist and I'm trying to generate a little bit more video before I put it up. The AI video generator has limits per day so I only get so many turns at the wheel 🤣

Thanks!!

Yeah, both at home and in the car! We need it in Spotify first tho!

Glad you enjoyed it

Thanks

Thanks! I hope to make a playlist about governance and strategy. Great videos on your channel too!