Code To The Moon @CodeToTheMoons

Software development YouTube channel specializing in Rust and LLMs. youtube.com/channel/UCjREV… Joined October 2022

Tweets

591
Followers

993
Following

194
Likes

853

Code To The Moon @CodeToTheMoons

4 weeks ago

At some point, a religious group will form based on the premise of divine revelation from a specific large language model. Calling it now.

0 0 2 171 0

Code To The Moon @CodeToTheMoons

4 months ago

😱

4 0 15 638 0

Download Image

Code To The Moon @CodeToTheMoons

5 months ago

My take on the current language model API market after a few days of research, and what I ultimately chose for my application. Did I miss anything notable? All of these have OpenAI compatible APIs, though don't necessarily support all of the same features 👇 @CerebrasSystems -…

0 0 1 247 2

Code To The Moon @CodeToTheMoons

5 months ago

Definitely agree with the general sentiment here 😎 But is drawing on prompt history the best way to understand a user?

Sam Altman @sama

5 months ago

Definitely agree with the general sentiment here 😎 But is drawing on prompt history the best way to understand a user?

2K 2K 31K 2.9M 3K

0 0 1 238 0

Code To The Moon @CodeToTheMoons

5 months ago

wow. Say what you like about the "vibes" of Llama 4 (for its cost I think it is actually pretty solid), but any way you look at it this is extremely impressive

Cerebras @CerebrasSystems

5 months ago

wow. Say what you like about the "vibes" of Llama 4 (for its cost I think it is actually pretty solid), but any way you look at it this is extremely impressive

23 49 446 48K 109

Download Image

0 0 2 306 0

Code To The Moon @CodeToTheMoons

5 months ago

Emacs is an amazing piece of software. But it’s been 3-4 months since I last used it, and I don’t think it’s the best choice for me anymore 💔 been loving Cursor + vim motions, but would also like to get back to a terminal based editor. What should I switch to?

4 0 10 597 0

Code To The Moon @CodeToTheMoons

5 months ago

Just came across @CerebrasSystems, an inference API provider that is orders of magnitude faster than the competition. Tested it myself and is appears to be the real deal. Near instantaneous inference even with large SOTA models. What am I missing here? What's the catch? Also…

1 0 0 178 0

Code To The Moon @CodeToTheMoons

5 months ago

what is the go-to LLM leaderboard these days? the HF leaderboard now says "archived" and does not include Gemma 3, nor does it appear to have a link to a successor...

0 1 1 232 0

Code To The Moon @CodeToTheMoons

5 months ago

Home security system has been hijacked by vibe coders

0 0 1 217 0

Download Image

Code To The Moon @CodeToTheMoons

5 months ago

In the night sky we see star constellations that look vaguely like familiar objects. But our perspective from Earth is just one in a galaxy of possible perspectives. What if there are interesting constellations that can be seen from vantage points elsewhere in the galaxy? Take…

0 0 2 196 0

Code To The Moon @CodeToTheMoons

5 months ago

Amazon Nova LLMs are actually really good. Even the micro and lite flavors are fast, inexpensive and accurate. I don't know how they fare in the standard benchmarks, but for my current use case they perform far better than the 1B, 4B and 12B flavors of Google's new Gemma 3 models

0 0 3 246 0

Code To The Moon @CodeToTheMoons

5 months ago

I may have transitioned away from Emacs and I just caught myself using a regular keyboard instead of split columnar. Appreciate your support during this difficult time.

1 0 12 340 0

Code To The Moon @CodeToTheMoons

6 months ago

Are there any widely lauded works of fiction that are known to be completely or mostly AI generated?

0 0 1 247 0

Code To The Moon @CodeToTheMoons

6 months ago

I managed to accrue $150 of AWS Bedrock charges simply by testing my application with 1 user over the course of about 1 hour. Self hosting is looking like a pretty compelling option at this point...

3 0 10 457 0

Code To The Moon @CodeToTheMoons

6 months ago

Cursor Agent mode is absolutely incredible. Super helpful but just fallible enough for me to avoid an existential crisis as a human developer. I haven’t tried Devin but I can’t shake the feeling that the “separate from your IDE” approach isn’t going to work until accuracy is…

4 0 8 655 1

Code To The Moon @CodeToTheMoons

7 months ago

I am stuck on page 3 of the DeepSeek paper. I cannot fully wrap my head around how V3 and R1 are related. It almost seems like they are interdependent on each other. @Sentdex @AIExplainedYT or anybody else help??