Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
peab
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
211.
▲
A Field Guide to Rapidly Improving AI Products
(oreilly.com)
2 points
by
peab
2y ago
|
0 comments
212.
▲
by
peab
2y ago
Thanks for the thoughtful reply. >In retrospect, I think it's possible I took all of this the wrong way. But even after reading the post a 2nd time, I can still read it both ways. I think what you're saying resonates with me, a
213.
▲
by
peab
2y ago
I'm definitely considering open sourcing it yeah! I'll post a follow up if/when I do.
214.
▲
by
peab
2y ago
I think code is purely functional - there might be a style component to it, but mostly what I care about is "does it do the thing I want it to do". Writing, on the other hand, is much more about expressing one's inner self. I
215.
▲
by
peab
2y ago
Yeah, reading through the top comments here I can't help but come to the conclusion that they didn't read the actual post!
216.
▲
by
peab
2y ago
I definitely agree with part of what you're saying. The AI tools make it much easier to become disconnected from what you're writing. That being said, did you read this post in particular? I explicitely say "ChatGPT and Claud
217.
▲
by
peab
2y ago
Yeah, i haven't seen a reason why we can't just use REST. Like, auth is already figured out. The LLMS already have the knowledge of how to call APIs too!
218.
▲
Like cursor, but for blogging: a weekend project
(maximepeabody.com)
38 points
by
peab
2y ago
|
54 comments
219.
▲
Building an AI-Assisted Blog Editor (it gets meta)
(maximepeabody.com)
2 points
by
peab
2y ago
|
0 comments
220.
▲
by
peab
2y ago
I hate to break it to you, but this sounds like pretty normal human social behavior
221.
▲
by
peab
2y ago
any API can be modeled as JSON in, JSON out, which you can pass to the system prompt at design time or at runtime, no?
222.
▲
by
peab
2y ago
Ah, that's another good use case. RIP to whatever company doesn't allow their workers to use the LLMs though
223.
▲
Local LLM Models: Are They Useful?
(maximepeabody.com)
1 points
by
peab
2y ago
|
4 comments
224.
▲
Show HN: Image Pop Art Tool
(imagepop.maximepeabody.com)
1 points
by
peab
2y ago
|
0 comments
225.
▲
A few thoughts on being a Generative AI Engineer
(maximepeabody.com)
2 points
by
peab
2y ago
|
0 comments
226.
▲
From Hype to Reality: What It Means to Be a GenAI Engineer
(maximepeabody.com)
2 points
by
peab
2y ago
|
0 comments
227.
▲
by
peab
2y ago
Fair. I promise I'm not collecting emails. TBH I don't think the hackernews crowd is the right target for a Tarot App lol. I just thought this was adjacent to it and would be interesting.
228.
▲
by
peab
2y ago
That's sort of the point. Those personality tests are similar to astrology - they're mostly not accurate, but they hold a tiny bit of truth. The question is, can that tiny bit of truth be useful? I posit that it can.
229.
▲
by
peab
2y ago
that's one of the first 10 links that show up on google when you search for "gpt4 training size". If it is AI generated slop, that's unfortunate. I probably didn't spend enough time looking for sources for the size
230.
▲
by
peab
2y ago
I knew someone would say that lol. The website itself has google auth, but for this feature I promise I'm not storing the emails. I know the hacker news crowd haha
231.
▲
What do LLMs know about your email address?
(youtarot.app)
3 points
by
peab
2y ago
|
2 comments
232.
▲
by
peab
2y ago
Here's a fun test to see what LLMs know about your email address. While most model providers put efforts into removing PII, there's still some information that can be inferred from your email, whether you realize it or not. It
233.
▲
What Your Email Address Reveals About You: LLMs and Digital Footprints
(maximepeabody.com)
24 points
by
peab
2y ago
|
16 comments
234.
▲
by
peab
2y ago
Oh good catch. Singsong should be in the infilling section. There's a Chinese lab model that does vocal synth but i forget the name of it!
235.
▲
by
peab
2y ago
Yes. I anticipate that the open source models will pave the way for that, just like we have in painting with stable diffusion. Fundamentally, a song can be represented as a 2d image without any loss
236.
▲
by
peab
2y ago
From the artist perspective, this is correct. But there are lots of applications for music which parallel the applications of ai generated images - things that are more commercial in nature. The media is functional, for use cases such as co
237.
▲
by
peab
2y ago
Yes well said. Distribution networks are hard to disrupt
238.
▲
Music Generation AI Models
(maximepeabody.com)
39 points
by
peab
2y ago
|
55 comments
239.
▲
Show HN: AI-Powered Audio Sample Generator (VST/AU)
(text-to-sample.com)
3 points
by
peab
2y ago
|
0 comments
240.
▲
by
peab
2y ago
Even if it's cultural, it can be fixed. Culture can change and can be changed by choice
More ›