Andrei Potekhin

Hola. Could you elaborate how do you actually generate svgs? I see two ways. You train a model to actually generate vector paths. Or you generate a raster image first and then vectorize it (as you would do in eg Illustrator). Is it one of these or something different? I tried generating svgs with ChatGPT about 2 years ago, and it did a fair job. Was able to make separate triangles for the...

Don't know how many times you've practiced the demo with your cat, but the timing was perfect! Cute product, congrats on the launch.

Looks nice. One note. Somehow, the synthetic voice in the video doesn't allow me to trust the product fully. I think it's essential to build trust especially when people share their files. So, I would avoid using generated voice and use a video with a real voice instead.

Slick design, folks! It seems there are a few tiny ones to fix in the mockups on this page. 3rd static mock: one chip is cropped form left and right, and the search placeholder says "Ask a anything" shouldn't it be "Ask anything?".

Hola. Conversational learning sounds very promising. Congrats! A few notes: - Why does the placeholder for the response say "Write something"? I think a more motivating message can encourage learners to put more effort into reflecting on the content. If the message is "Write something", learners may just write a one-word response. What if it said smth like "What are your insights from this...