R to @rauchg:

R to @levelsio: Keep it up!

R to @ammaar: Great thread! Sorry for the noob question, but where/how do I use the model once it is trained to generated the images?

R to @swyx: iminthearena.dev

R to @levelsio: Using SDXL now?

R to @levelsio: The world of app development === waiting for Apple

R to @ParkerRex: What's a sundial? No just brain

R to @levelsio: Hey, let us help you.

R to @duborges: You get 2 sec gen time on your own server? What specs?

R to @levelsio: Very slow, should take 2 secs per image.

R to @levelsio: What's your influencer outreach strategy? (your offer to them)

R to @JesseLempiainen: how is it different from HuggingFace?

R to @levelsio: how do u track ur goals? sundial + stickie ?

R to @levelsio: Found repicate because of your previous tweet. Have been on it for 2 months now, works insanely well!

R to @duborges: Ah ok so yes that's fast no?

R to @usamaClick: I just ran another one via their dashboard: Generated in 11.98 seconds. 12 x 4 = 48 seconds.

RT by @BDelipetrev: I record in segments of 15-45 minutes using OBS, after a segment is done, I move it to premiere pro, watch in x1.25 speed, adjust the sound, maybe add some graphics / post commentary etc. and render the segment. This keeps my mind at at ease that video is not corrupted. I name the rendered file 01_SPOTIFY_INTRO for example, and increase the number of each segment. After I finish I have about 15-20 segments, and since I named them 01,02,03,04... I can easily add them all to premiere pro again in correct order, and since I already watched each of them I don't have to double-check the 6, 8 ,10 hour video that I have in the end. It is probably a very not optimized process but it keeps my mind at ease with these long videos :)