Hacker News: humblyCrazyhttps://news.ycombinator.com/user?id=humblyCrazyHacker News RSShttps://hnrss.org/hnrss v2.1.1Thu, 25 Jun 2026 02:49:56 +0000<![CDATA[New comment by humblyCrazy in "Computer use in Gemini 3.5 Flash"]]>I looked at their demo and it does not

]]>Wed, 24 Jun 2026 18:23:50 +0000https://news.ycombinator.com/item?id=48663809humblyCrazyhttps://news.ycombinator.com/item?id=48663809https://news.ycombinator.com/item?id=48663809<![CDATA[New comment by humblyCrazy in "Qwen3-Next"]]>How is MTP different from Medusa heads? Also does this mean this model comes "natively" with speculative decoding - meaning if I use this model in vllm, it's throughput should be higher because it is already doing MTP so it should be able to take advantages of speculative decoding?

]]>Fri, 12 Sep 2025 14:04:38 +0000https://news.ycombinator.com/item?id=45222292humblyCrazyhttps://news.ycombinator.com/item?id=45222292https://news.ycombinator.com/item?id=45222292<![CDATA[New comment by humblyCrazy in "The Agent2Agent Protocol (A2A)"]]>i dont understand how it is different from mcp. The blog just says "A2A is an open protocol that complements Anthropic's Model Context Protocol (MCP), which provides helpful tools and context to agents." There is no example or anything on how does it complement it

]]>Wed, 09 Apr 2025 14:26:01 +0000https://news.ycombinator.com/item?id=43632529humblyCrazyhttps://news.ycombinator.com/item?id=43632529https://news.ycombinator.com/item?id=43632529