TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

100M Token Context Windows

94 pointsby gklitt9 months ago

8 comments

shazami9 months ago
FYI wouldn't interview here. Got rejected after a 30 minute behavioral screen after spending 8 hours on an unpaid take-home.
评论 #41406384 未加载
评论 #41399278 未加载
dinobones9 months ago
Long context windows are IMO, “AGI enough.”<p>100M context window means it can probably store everything you’ve ever told it for years.<p>Couple this with multimodal capabilities, like a robot encoding vision and audio into tokens, you can get autonomous assistants than learn your house&#x2F;habits&#x2F;chores really quickly.
评论 #41397412 未加载
评论 #41395245 未加载
评论 #41395642 未加载
smusamashah9 months ago
It should be benchmarked against something like RULER[1]<p>1: <a href="https:&#x2F;&#x2F;github.com&#x2F;hsiehjackson&#x2F;RULER">https:&#x2F;&#x2F;github.com&#x2F;hsiehjackson&#x2F;RULER</a> (RULER: What’s the Real Context Size of Your Long-Context Language Models)
评论 #41394574 未加载
评论 #41394468 未加载
fsndz9 months ago
Context windows are becoming larger and larger, and I anticipate more research focusing on this trend. Could this signal the eventual demise of RAG? Only time will tell. I recently experimented with RAG and the limitations are often surprising (<a href="https:&#x2F;&#x2F;www.lycee.ai&#x2F;blog&#x2F;rag-fastapi-postgresql-pgvector" rel="nofollow">https:&#x2F;&#x2F;www.lycee.ai&#x2F;blog&#x2F;rag-fastapi-postgresql-pgvector</a>). I wonder if we will see some of the same limitations for long context LLM. In context learning is probably a form of semantic &#x2F; lexical cues based arithmetic.
Sakos9 months ago
I was wondering how they could afford 8000 H100’s, but I guess I accidentally skipped over this part:<p>&gt; We’ve raised a total of $465M, including a recent investment of $320 million from new investors Eric Schmidt, Jane Street, Sequoia, Atlassian, among others, and existing investors Nat Friedman &amp; Daniel Gross, Elad Gil, and CapitalG.<p>Yeah, I guess that&#x27;d do it. Who are these people and how&#x27;d they convince them to invest that much?
评论 #41395207 未加载
评论 #41394955 未加载
anonzzzies9 months ago
What is the state of art on context on open models? Magic won&#x27;t be open I guess after getting 500m in VC money.
samber9 months ago
Based on Mamba ?
htrp9 months ago
does anyone have a detailed tech breakdown of these guys? not quite sure how their LTM architecture works.
评论 #41395903 未加载