owenpalmer

12 ms·

91.
by owenpalmer 2y ago
> find a bunch of Lego fan gamers and hire them to make experimental games I love this idea, imagine if Lego had an open game creation platform. So much potential for idea generation.
92.
by owenpalmer 2y ago
Isn't this just a prompt cookbook? I was expecting a model specifically trained to teach students faster or more effectively.
93.
by owenpalmer 2y ago
I meant VTT, not TTS
94.
by owenpalmer 2y ago
I'm not sure encouragment itself is the performance enhancer, it's more that you're communicating that the model has the right "vibe" of what your end goal is.
95.
by owenpalmer 2y ago
This is actually a great approach. Essentially you're using time travel to prevent misunderstandings, which prevents the context from getting clogged up with garbage.
96.
by owenpalmer 2y ago
A few options. 1. Use TTS and have an LLM clean it up. 2. Use a collection of prompt templates.
97.
by owenpalmer 2y ago
Synthetic data will play I big role, yes. There's other challenges though, like how verbal descriptions of objects would affect their spatial behavior. Building a generalized simulator that combines those modalities is hard. In this pa
98.
by owenpalmer 2y ago
> All models exhibited limitations in spatial planning when constructing multi-section factories. Common failures included placing entities too close together, not allocating space for connections, or incorrect inserter placement It make
99.
by owenpalmer 2y ago
The one that Claude created was a legitimately fun game! If it implemented boids similar to o3-mini, it would be even better. Slap some sprites on it and put it on steam!
100.
by owenpalmer 2y ago
Which portion of the text gave you that impression?
101.
by owenpalmer 2y ago
Thanks for the link! Good to know someone is working on something similar.
102.
by owenpalmer 2y ago
I've been searching relentlessly for something like this! I wonder why it's been so hard to find... is it the Chinese? In any case, thanks for sharing.
103.
by owenpalmer 2y ago
This is incredibly exciting. I've been pondering/experimenting on a hobby project that makes reading papers and textbooks easier and more effective. Unfortunately the OCR and figure extraction technology just wasn't there yet
104.
by owenpalmer 2y ago
I think you're missing a key point, which is that by extensively reading the R1 outputs, I'm able to observe how R1 thinks about things, which I can then replicate. There are good ways and bad ways to think aloud, R1 just gave me
105.
by owenpalmer 2y ago
> four key cognitive behaviors -- verification, backtracking, subgoal setting, and backward chaining -- that both expert human problem solvers and successful language models employ. As we make AI better, perhaps we'll inadvertently
106.
by owenpalmer 2y ago
> The stop in sharing data was “due to funding constraints that have caused the Department to turn off the underlying network” What is the actual recurring cost of broadcasting this data? The sensor and network infrastructure are presuma
107.
by owenpalmer 2y ago
I remember being a little kid learning Blender, rooting for it to become the industry standard. It's amazing to see how much the project has grown.
108.
by owenpalmer 2y ago
Chobani made a really beautiful ad featuring the solarpunk aesthetic: https://youtu.be/z-Ng5ZvrDm4?si=BEmNr2kaBblgI64v People hate on it for different reasons, but I like the vision/aesthetic they're going for. No
109.
by owenpalmer 2y ago
My biology professor provides basically zero feedback on his student's understanding of the material. There are very few practices questions to prepare for exams, which are worth 40% of your grade. I had an LLM write some python that e
110.
by owenpalmer 2y ago
> perhaps you are the exception that proves the rule As someone who worked directly with OP, I can confirm they are highly exceptional. Don't be too surprised when you encounter talented people on HN. I challenge you to be more inqu
111.
by owenpalmer 2y ago
The large binary size isn't because of the language. It's just really poor optimization.
112.
by owenpalmer 2y ago
I could imagine this with automatic TTS and a slideshow. Basically you'd have procedurally generated mini documentaries in a tiktok like format.
113.
by owenpalmer 2y ago
The first sentence of the article says "some payments" are being shut down.
114.
by owenpalmer 2y ago
> we will obviously deliver much better models It's not obvious Sam, that's the problem.
115.
by owenpalmer 2y ago
This sounds great! Is there any sample curriculum or a demo to try out? I love the idea of teaching maths using video games or simulations.
116.
What Everyone Gets Wrong about AI(youtube.com)
1 points by owenpalmer 2y ago|0 comments
117.
Creating Scenes in Godot with AI (Claude and Trellis)(youtube.com)
2 points by owenpalmer 2y ago|0 comments
118.
by owenpalmer 2y ago
> Perhaps a simple gimmick like ducktape over the mouth? He has a beard, the removal process would be extremely painful.
119.
by owenpalmer 2y ago
LLMs are deterministic. It's just that the random seed is hidden from you, but is still an input.
120.
by owenpalmer 2y ago
Have you tried RAG on the docs?

More ›