HackerTrans
TopNewTrendsCommentsPastAskShowJobs

xianshou

no profile record

Submissions

Merge and Conquer: Evolutionarily Optimizing AI for 2048

arxiv.org
1 points·by xianshou·9 माह पहले·0 comments

Stuck in the Matrix: Probing Spatial Reasoning in Large Language Models

arxiv.org
1 points·by xianshou·9 माह पहले·0 comments

Reflection AI Raises $2B to Build "American DeepSeek"

nytimes.com
9 points·by xianshou·9 माह पहले·2 comments

Nvidia-backed Reflection AI raising at $5.5B valuation

reuters.com
2 points·by xianshou·9 माह पहले·1 comments

comments

xianshou
·2 माह पहले·discuss
A lovely example of a study that is both obviously true and misses the point.

Music with lyrics directly interferes with any task that has a verbal component, and the worse you are at multitasking, the worse the interference. Despite being terrible at multitasking, I still listen to music with lyrics. Why? Principally because the alternative, hearing all the conversations in my immediate vicinity, is usually both more distracting and less pleasant. But there are also auxiliary benefits, such as an increase in "work stamina" and a passive signal to coworkers to interrupt only if it's important.

Now, I could listen to lo-fi all day, or three-hour soundtracks on Youtube, and sometimes do, but it gets boring pretty fast!

Anyway: obviously true, still worth it because the alternative is worse.

(By the way, other mitigating strategies: listening to music in a language you don't understand, or listening to lyrics so familiar you can screen them out. My top Spotify songs all get played several hundred times a year.)
xianshou
·3 माह पहले·discuss
Even as someone extremely firmly on the other side of the AI debate, I must appreciate the craft.

Now, to give Claude the steganogravy skill...
xianshou
·3 माह पहले·discuss
From the file: "Answer is always line 1. Reasoning comes after, never before."

LLMs are autoregressive (filling in the completion of what came before), so you'd better have thinking mode on or the "reasoning" is pure confirmation bias seeded by the answer that gets locked in via the first output tokens.
xianshou
·3 माह पहले·discuss
I appreciate not having to read this guy again.
xianshou
·5 माह पहले·discuss
Great work! Why no benchmarks though?
xianshou
·5 माह पहले·discuss
Nice! 5 bucks says you can swap this in for your average software kanban and it does a better job.
xianshou
·5 माह पहले·discuss
Safer than clawdbot/moltbot, I'll bet.
xianshou
·6 माह पहले·discuss
Incidentally, Chroma also produced the single best study on long-context degradation that I've come across:

https://research.trychroma.com/context-rot

Before that, I cited nolima (https://www.reddit.com/r/LocalLLaMA/comments/1io3hn2/nolima_...) constantly to illustrate how difficult tasks involving reasoning or multi-step information gathering degraded much faster than the needle-in-haystack benchmarks cited by the major labs. Now Chroma is the first stop. Nice job on the research!