<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: didibus</title><link>https://news.ycombinator.com/user?id=didibus</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sat, 30 May 2026 19:23:15 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=didibus" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>I still didn't follow, you mean to pipe things between tool calls? Like if you want to query something and then update another without the intermediate getting brought in context?</p>
]]></description><pubDate>Sat, 30 May 2026 07:35:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=48333664</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48333664</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48333664</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>> now I just write tools for agents to consume<p>What do you mean? Tool is a pretty generic concept.</p>
]]></description><pubDate>Sat, 30 May 2026 05:45:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=48333003</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48333003</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48333003</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>I think if you deny a tool, it won't be loaded in context at all ever, even it's name and description won't be loaded.</p>
]]></description><pubDate>Sat, 30 May 2026 05:40:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332987</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332987</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332987</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>Deferred cli/skill loading is also not part of CLIs or skills, it's all about how the coding agent/harness is implemented.</p>
]]></description><pubDate>Sat, 30 May 2026 05:38:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332974</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332974</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332974</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>Skill still needs to be loaded in context, what would it change?</p>
]]></description><pubDate>Sat, 30 May 2026 05:37:58 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332968</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332968</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332968</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>Not just that, but they retracted this:<p>> Update: Since these measurements were taken, Claude Code has rolled out Tool Search with Deferred Loading, which loads MCP tool schemas on-demand and reduces context usage by 85%+. The context bloat described in Problem 1 is largely addressed for users on current Claude Code versions. The performance, debugging, and architectural arguments below still apply.<p>Because Claude Code only loads the tools it needs now, so context bloat is pretty much solved for MCPs.</p>
]]></description><pubDate>Sat, 30 May 2026 05:36:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332962</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332962</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332962</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>You do know MCPs are loaded on demand same as skills now right? The only place where sometimes it still uses too much context is if you have too many MCPs (same issue with skills) or some MCP is poorly designed and responds with huge description or MCP calls respond with way too much info, but skills can have this issue as well.</p>
]]></description><pubDate>Sat, 30 May 2026 05:32:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332943</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332943</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332943</guid></item><item><title><![CDATA[New comment by didibus in "MCP is dead?"]]></title><description><![CDATA[
<p>A CLI needs to work on windows, mac, linux, android, iOS, etc. And it still needs some backend APIs to call. So creating one is a lot more work than just making an MCP.</p>
]]></description><pubDate>Sat, 30 May 2026 05:26:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=48332897</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48332897</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48332897</guid></item><item><title><![CDATA[New comment by didibus in "She can mentally time travel. Why did everyone think she was lying?"]]></title><description><![CDATA[
<p>I don't think it's a condition that is being debated much, it seems to have been pretty well studied, though it doesn't seem we really understand how it works or why it happens. At least based on the Wikipedia page.</p>
]]></description><pubDate>Tue, 26 May 2026 07:07:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=48276161</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48276161</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48276161</guid></item><item><title><![CDATA[New comment by didibus in "She Can Mentally Time Travel. Why Did Everyone Think She Was Lying?"]]></title><description><![CDATA[
<p>If you read the article, it's because she has perfect memory recall, so they say it's as-if she can travel back in time, since she remembers everything vividly in great details perfectly.<p>Apparently the condition is called hyperthymesia.</p>
]]></description><pubDate>Tue, 26 May 2026 02:54:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=48274451</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48274451</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48274451</guid></item><item><title><![CDATA[New comment by didibus in "Why is almost everyone right-handed? A new study connects it to bipedalism"]]></title><description><![CDATA[
<p>Interesting, but it confuses me a bit.<p>Ambidextrous means: "that the person has no marked preference for the use of the right or left hand".<p>Unlike what is normally assumed, it doesn't mean you've mastered both hands, it doesn't imply you can write equally well with the left or the right, but that you could have just as easily learned either. You had no marked preference when you picked up the activity initially.<p>Cross-dominance seem to imply there was a marked preference, but that it is the left or right hand differs per activity.<p>That said, I wonder if they're really the same. Often people say ambidextrous as you are equally good with both hands, which would always require practice on both side. But maybe cross-dominant people can equally learn?<p>Under that reading, cross-dominance could be what ambidexterity actually looks like in practice. If you genuinely have no preference, there's no reason you'd consistently land on the same hand across all activities, you'd be influenced by context, who taught you, which hand was free, etc.<p>Have you ever tried your other hand at activities? And are you surprisingly good with it, even if not as good? I tend to be better using my other hand at most things than say a fully right handed person, but never as good as the hand I've been using for that activity consistently.</p>
]]></description><pubDate>Wed, 20 May 2026 06:04:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=48203682</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48203682</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48203682</guid></item><item><title><![CDATA[New comment by didibus in "Tesla's lithium refinery discharges 231,000 gallons of polluted wastewater a day"]]></title><description><![CDATA[
<p>Personally, I find all these "do they have the necessary approvals or not" discussions are besides the point that matters. For business, those are red tape, wasted time, and unnecessary bureaucracy. For residents and citizens, what permits were approved or not don't matter, but if we tolerate the pollution cost that results, no matter if it was legally approved or not.<p>I'd rather an article that argues about if this pollution cost that is being externalized to Texans to pay, justified and a net win for them, and if it is, than what's holding up the permits, and if not, then why is this permitted at all, even if partially.</p>
]]></description><pubDate>Wed, 20 May 2026 00:04:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=48201336</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48201336</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48201336</guid></item><item><title><![CDATA[New comment by didibus in "Why is almost everyone right-handed? A new study connects it to bipedalism"]]></title><description><![CDATA[
<p>> why handed vs ambidextrous<p>Did it even explain that? I'm ambidextrous, I have no handedness bias, so whichever I pick up to first learn something is the hand I use. So I'm a mix of left-handed and right-handed depending on the task. And yet I didn't really understand why that's odd because of my bipedalism?</p>
]]></description><pubDate>Tue, 19 May 2026 22:53:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=48200761</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48200761</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48200761</guid></item><item><title><![CDATA[New comment by didibus in "Mercurial, 20 years and counting: how are we still alive and kicking? [video]"]]></title><description><![CDATA[
<p>I often want to "save" but not have a comment, and not ready to make it a clean commit that I want a comment on. That's when I stage, then I can see the diff and revert still. But ya, maybe I could adapt to not worrying about having a million commits instead of clean ones at points that make sense with good comments.</p>
]]></description><pubDate>Mon, 18 May 2026 04:32:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=48175608</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48175608</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48175608</guid></item><item><title><![CDATA[New comment by didibus in "It is time to give up the dualism introduced by the debate on consciousness"]]></title><description><![CDATA[
<p>I suspect more things are conscious than we tend to assume. I would assume some level of intelligence requires a review/assessment process, something to evaluate what happened, what is good or bad, what should we have done instead, how can we do it better next time. This self-assessment becomes our experience of consciousness. Of course it feels incredible, unreal, like those feelings overwhelm us, because we are this function, and optimizing for those feelings is our function.</p>
]]></description><pubDate>Mon, 18 May 2026 04:04:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=48175455</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48175455</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48175455</guid></item><item><title><![CDATA[New comment by didibus in "Claude for Small Business"]]></title><description><![CDATA[
<p>A plugin is just a bundle of MCPs, skills and templated prompts.<p>A skill cannot provide MCPs and can't provide custom template prompts, each skill is it's own slash command.<p>A plugin you can define N number of custom slash commands, and you can define MCPs as well as skills. So it bundles like all the things together.<p>By installing a plugin, you are basically installing a bunch of MCPs, skills and custom slash command prompts.</p>
]]></description><pubDate>Thu, 14 May 2026 05:02:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=48131285</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48131285</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48131285</guid></item><item><title><![CDATA[New comment by didibus in "Show HN: I made a Clojure-like language in Go, boots in 7ms"]]></title><description><![CDATA[
<p>Cool, I think at this point, since there's a few Clojure Go dialect, what I'd like to understand most is which one is fully hosted on Go, has as good interop as JVM Clojure has with Java, where I can both use what I make from Go, and also use Go from Clojure, and even create mixed projects would be nice. And also, truly details the differences, beyond interrop, what is not the same and differs.</p>
]]></description><pubDate>Sun, 10 May 2026 19:36:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=48087074</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48087074</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48087074</guid></item><item><title><![CDATA[New comment by didibus in "Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc"]]></title><description><![CDATA[
<p>Interesting, but why not then use an even stricter language? Say Idris, ATS, Lean or F* ?</p>
]]></description><pubDate>Sun, 10 May 2026 19:15:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=48086898</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48086898</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48086898</guid></item><item><title><![CDATA[New comment by didibus in "Cloudflare to cut about 20% of its workforce"]]></title><description><![CDATA[
<p>It kind of depends what we mean. If you're conservatively in the market, invested in the aggregate economy, diversified, and what not, yes, but if you're taking bets on a smaller number of companies you can just lose your money full on. Not every single company recovers from a recession.<p>That's why if you are a business, the risk of a recession is a real threat. Someone will recapture your market once the recession is over, but will you?<p>That also means people will lose their jobs, price of goods will rise, the pressure to need to dip into one's savings will increase, forcing many people into cashing out at the worse possible time. If you are someone with that risk, as an individual, a recession is a real threat as well, and you might want to reduce your market exposure beforehand.</p>
]]></description><pubDate>Fri, 08 May 2026 22:08:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48069378</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48069378</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48069378</guid></item><item><title><![CDATA[New comment by didibus in "Cloudflare to cut about 20% of its workforce"]]></title><description><![CDATA[
<p>You can't really compare them to Microsoft, Oracle, or Meta. Those companies aren't cutting costs because AI replaced their own employees. They're pouring money into AI infrastructure and models because they want to sell that capacity to others.<p>Their thinking is more: instead of funding another internal product team, they can redirect that payroll spend into more AI compute and models they hope to monetize.<p>I don't believe CloudFlare is doing that, though they might, they could be needing to spend in Edge AI compute and what not, building out that infra isn't free, so they might need to find places the cash will come from.</p>
]]></description><pubDate>Fri, 08 May 2026 05:59:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=48059156</link><dc:creator>didibus</dc:creator><comments>https://news.ycombinator.com/item?id=48059156</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48059156</guid></item></channel></rss>