3 ms·
GPT-3 is already better than most blog comments, many bloggers, and most online text porn. Maybe most ad copy. Many pundits. Some college professors. It has no
by Animats 6y ago
GPT-3 is already better than most blog comments, many bloggers, and most online text porn. Maybe most ad copy. Many pundits. Some college professors.
It has no clue what it's talking about, of course. It's terrible at fact-based material, because what comes out looks plausible, but contains made-up facts. Some facts will be correct, because they were mentioned a few times in the training database. Others are just plausible words strung together.
What it really demonstrates is that much of literature has less behind it than was previously supposed.
- FeepingCreature 6y agoI expect GPT-3 to be amazing for people who want to write stories or articles but are really terrible at one specific aspect of the writing process.
- sbierwagen 6y agohttps://slideplayer.com/slide/6081467/18/images/12/The+scale+of+intelligent+minds%3A+a+parochial+view..jpg https://slideplayer.com/slide/6081467/18/images/12/The+scale... It depends on if you think writing ability is more like the top graph, or the bottom graph. If it's like the top graph, humans have plenty of time to adopt AI tools, and incorporate them into the writing process. If it's like the bottom graph, then GPT-4 comes out in 12 months and is human equivalent, and GPT-5 comes out a year after that and is a better writer than any human could hope to be.
- Animats 6y agoProgress so far is that GPT-1 could produce a coherent sentence but stopped making sense after a few lines. GPT-2 could produce a coherent paragraph, but several paragraphs strung together made no sense. GPT-3 is coherent for about a page or screen, but not a chapter. A page is good enough for the clickbait industry.
- bpodgursky 6y agoI have a friend who uses GPT-3 to generate startup marketing copy and customer-facing emails. He just prompts with the company description and "The email which is most effective at getting customers to sign up for <startup> is:" It reads as good or better than a human would do. Maybe not better than the BEST human marketer would do, but far better than average, and instant + free(ish).
- malloryerik 6y ago> What it really demonstrates is that much of literature has less behind it than was previously supposed. I'm uncomfortable with this statement because GPT-3 is in a sense copying or sourcing a massive corpus of human literature to use as the paint on its brush. If it can make a sentence that sounds Shakespearean after performing a statistical analysis of all sentences ever written by Shakespeare and using fancy averages and weightings of those sentences, how does that diminish the value of Shakespeare? In a wider sense, isn't that what's happening, but instead of using every sentence written by Shakespeare the goal is to use every sentence ever written? That said, how GPT-3 and its descendants affect the market value of "literature" is... sorry I can't resist... a story yet to be written.