Etsy employee #3 or so here but haven’t worked there in more than a decade. Rob is a great guy, but I don’t think he could have grown Etsy the way it has. I’m sure some people will say that’s not a bad thing but my response is you probably wouldn’t know about Etsy if he stayed on.
I think on the whole, the new CEO has done more good than bad for the company. They’ve always had criticism of non handmade stuff being sold on there. I think they could do more to that end, and if the video is right that the new CEO is allowing non handmade stuff on there, I don’t agree with him on that. I haven’t seen that myself and I do still use the site. While he’s made other decisions I don’t agree with, encouraging sellers to do free shipping was a good move. Many buyers expect that thanks to Amazon. The fee increases while for sure had an impact on sellers bottom lines, don’t compare to what Amazon Handmade (if that still exists) and ebay charge (not to get into most other marketplaces like the app stores that charge 30%). The current CEO in my opinion understands Etsy way more than the other two they had after Rob was out.
Also in terms of Fred Wilson, she should have done a little more homework on him. He was one of the original investors. He understands Etsy. He’s also entitled to some return for making a very risky investment on 4 kids (they were like 20 when they started it). I haven’t spoken to Fred in some time so maybe he’s changed, but I doubt it.
Anyway, I don’t mean to be so negative about the video, but I also don’t think Etsy has lost its way as much as the video implies. Granted I am not a seller, just a user at this point.
Wait a second here… I skimmed the paper and GitHub and didn’t find an answer to a very important question: is this GPT3.5 or 4? There’s a huge difference in code quality between the two and either they made a giant accidental omission or they are being intentionally misleading. Please correct me if I missed where they specified that. I’m assuming they were using GPT3.5, so yeah those results would be as expected. On the HumanEval benchmark, GPT4 gets 67% and that goes up to 90% with reflexion prompting. GPT3.5 gets 48.1%, which is exactly what this paper is saying. (source).