I am a non-coding consumer. I needed some randomized synthetic data in Google forms for a class project for my students to be later analyze but I didn’t want to manually fill in 50 surveys. O1 set me up with every program I needed and wrote a script which randomly filled out the Google forms survey questions I created with also fill in answers. There was added weights within questions so more realistic responses. To top it off it randomized the seconds it waited before re-running the script loop to avoid Google triggering a safety response and shutting down the automation. It did take a tiny bit of troubleshooting back and forth but like I said I do not code so I would send back the error and o1 would work with me to troubleshoot. I realize it wasn’t me doing the work but I felt like a goddamn genius when it worked 😂.
o1 is quite good a solving larger and more subtle problems. I have a class of problem that I basically reserve o1 and every time I get new tokens each week I burn through them in a day or 2 with problems that other models are failing at.
the thing with context is, I've been using wind surf and it doesn't have any of the context issues anthropic or chatgpt have even though it is running their models. so just the wrapper is enough to get their models to hold context properly.
Really looking forward to code mode!!! Also, I've been very impressed with the Google experimental model - really feels like Google is starting to get their stuff together. Veo 2 is wild.
Stargate: I loved the part where apparently SoftBank proposed to invest 100bn and Trump said "you should go for 500bn", and now that's it. I mean totally serious crowd. 😂
I am a non-coding consumer. I needed some randomized synthetic data in Google forms for a class project for my students to be later analyze but I didn’t want to manually fill in 50 surveys. O1 set me up with every program I needed and wrote a script which randomly filled out the Google forms survey questions I created with also fill in answers. There was added weights within questions so more realistic responses. To top it off it randomized the seconds it waited before re-running the script loop to avoid Google triggering a safety response and shutting down the automation. It did take a tiny bit of troubleshooting back and forth but like I said I do not code so I would send back the error and o1 would work with me to troubleshoot.
I realize it wasn’t me doing the work but I felt like a goddamn genius when it worked 😂.
I love Fridays! Great episode boys.
thanks I will try the code preview option:) - for data visualisation - it looks great
Great show, woefully under subbed
Is it Friday already?
the fact that o1 can now do images is even more proof that its just a wrapper for GPT4o
The guys need to rename the show to: The AI Comedy Hour and add a musical guest.
31:19 we need tooling stupid 😂 preach brother!
o1 is quite good a solving larger and more subtle problems. I have a class of problem that I basically reserve o1 and every time I get new tokens each week I burn through them in a day or 2 with problems that other models are failing at.
the thing with context is, I've been using wind surf and it doesn't have any of the context issues anthropic or chatgpt have even though it is running their models. so just the wrapper is enough to get their models to hold context properly.
Bout bloody time
O1 It's great when scientists use it for mathematics.
1:04:41 ARM is ARM. 🙂
Really looking forward to code mode!!!
Also, I've been very impressed with the Google experimental model - really feels like Google is starting to get their stuff together. Veo 2 is wild.
This Week In "Claude is the best" Podcast
im pissed hearing dario say hes giving up on consumer product. why the fuck would i pay $20 if i could just create my own app that calls their api?
Gemini 1206 is da bomb
I just don’t see the use cases for AI in the consumer space. But as a tool to help me at work I use it everyday.
Novelty value
Stargate: I loved the part where apparently SoftBank proposed to invest 100bn and Trump said "you should go for 500bn", and now that's it. I mean totally serious crowd. 😂
Males it up as he goes along lol
"grounded" 😂
Mrna. Oh dear.
I think you all are off base here or just not prompting it well, because o1 is actually so good as to be almost useful. it is definitely good.
Almost useful.. Perfect summary
The massive anthropic shills are back 👍
Does this guy sound exactly like Karp or is it me?
Didn't know fish could speak