danielcorin

6 ms·

1.
Stateful Agents and Basic Memory(danielcorin.com)
2 points by danielcorin 7mo ago|0 comments
2.
by danielcorin 1y ago
I’ve been building a Swift app [1], compatible with OpenAI APIs, easy model switching across providers, and with hotkeys for OS integration to capture text and images. It’s far more minimal than most other LLM frontends I’ve tried, but it’s
3.
Gemini 2.5 Flash default time to first token increased from 2.0(thoughteddies.com)
1 points by danielcorin 1y ago|0 comments
4.
by danielcorin 2y ago
More recently, `gemini-exp-1206` did quite well [1]. [1]: https://github.com/simonw/pelican-bicycle/blob/main/README.m...
5.
by danielcorin 2y ago
This is nice and has a very snappy UX. For the Vercel folks, how are you all thinking about this tool compared to offerings from Github[1] or Sourcegraph[2]? [1]: https://cs.github.com [2]: https://sourcegraph.com&#x2
6.
Cogno – Word Guessing Game(cogno.fun)
1 points by danielcorin 2y ago|1 comments
7.
by danielcorin 2y ago
Cogno is a word game that is the inverse of Semantle. You start with a word and the goal is to guess as many of the top 50ish semantically similar words as you can before you make 5 guesses outside the top list. You get points based on how
8.
Sonnet 3.5 can solve 40% of NYTimes Connections puzzles(danielcorin.com)
2 points by danielcorin 2y ago|0 comments
9.
VLMs Hallucinate(danielcorin.com)
2 points by danielcorin 2y ago|0 comments
10.
Protobuf-driven data extraction with language models(github.com)
2 points by danielcorin 2y ago|0 comments
11.
VLMs Aren't Blind(danielcorin.com)
2 points by danielcorin 2y ago|0 comments
12.
Evals: Unit Testing for Language Models(danielcorin.com)
1 points by danielcorin 2y ago|0 comments
13.
A few options for LLM result streaming with SSE(danielcorin.com)
1 points by danielcorin 3y ago|0 comments
14.
Simple sandboxed Python environment for LLM agent testing(danielcorin.com)
3 points by danielcorin 3y ago|0 comments
15.
(Unsuccessfully) Fine-tuning GPT to play "Connections"(danielcorin.com)
95 points by danielcorin 3y ago|50 comments
16.
by danielcorin 3y ago
https://blog.danielcorin.com/ I mostly write about LLMs (lately) or whatever I'm learning or working on
17.
Figuring out how to use LLMs in production(danielcorin.com)
2 points by danielcorin 3y ago|0 comments
18.
Shaping LLM Responses(danielcorin.com)
3 points by danielcorin 3y ago|0 comments
19.
Using GPT-3.5 to Quickly Generate and Run Shell Commands(danielcorin.com)
2 points by danielcorin 3y ago|0 comments
20.
Nix and Direnv(danielcorin.com)
1 points by danielcorin 3y ago|0 comments
21.
Beating Language Model Prompt Injection with Focus(danielcorin.com)
3 points by danielcorin 4y ago|0 comments
22.
by danielcorin 4y ago
I have some rough details on how I operated the AI tools here: https://adventure-of-penelope.vercel.app/credits . Unfortunately, I've lost the original log of the conversation.
23.
by danielcorin 4y ago
Thanks for the kind words
24.
by danielcorin 4y ago
Absolutely. This was recycled almost directly from some of the prompt material I gave while refining. I've learned more about "prompt engineering" (as it seems to be called) since and think it would be relatively easy to smoo
25.
by danielcorin 4y ago
I used very basic prompts including words like "cute" and "happy" and environment descriptions but not much beyond that (no art style or specific colors of the characters). I felt overall the character consistency has ro
26.
by danielcorin 4y ago
This piqued my interest. I might give it a shot, either with this or something new.
27.
by danielcorin 4y ago
~15 minutes with ChatGPT to turn the initial plot into a story with the tone and length I was looking for, an hour or two with Midjourney illustrating and a few more hours for the site from zero to up and running.
28.
by danielcorin 4y ago
I was going for more of a kid's book tone but I think I hadn't quite learned enough about the depth and nuance you can introduce into the output via prompting. In retrospect now, I agree it has that generic-AI tone.
29.
by danielcorin 4y ago
The porcupine has two mouths as well. I tried several times to smooth that one out but I could never quite reproduce the same look in the porcupine's eyes so I kept it.
30.
Show HN: I “wrote” a kid's book with ChatGPT and Midjourney(adventure-of-penelope.vercel.app)
417 points by danielcorin 4y ago|394 comments

More ›