<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: Reebz</title><link>https://news.ycombinator.com/user?id=Reebz</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 18 Nov 2026 03:58:48 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=Reebz" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by Reebz in "Ask HN: What's your current AI agentic setup?"]]></title><description><![CDATA[
<p>Little Linux server with: Orca IDE + Tailscale + 3x Claude code (builds) + Codex (reviews).<p>I run the Compound Engineering plugin as the primary work loop for both.<p>Orca is great, I can work anywhere.</p>
]]></description><pubDate>Fri, 06 Nov 2026 07:31:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=50389282</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=50389282</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50389282</guid></item><item><title><![CDATA[New comment by Reebz in "Tmux and Claude Code: Sessions, Send-Keys, and the Ceiling"]]></title><description><![CDATA[
<p>I went down this path over 6 months ago and had a variety of custom setups using tmux, cmux, and more. But since using Orca IDE it feels like a leap in usability and robustness.</p>
]]></description><pubDate>Mon, 19 Oct 2026 10:09:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=50150507</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=50150507</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50150507</guid></item><item><title><![CDATA[Karpathy-Rubin Spectrum]]></title><description><![CDATA[
<p>Article URL: <a href="https://alwaysdraft.com/karpathy-rubin-spectrum/">https://alwaysdraft.com/karpathy-rubin-spectrum/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49999147">https://news.ycombinator.com/item?id=49999147</a></p>
<p>Points: 5</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 07 Oct 2026 21:39:35 +0000</pubDate><link>https://alwaysdraft.com/karpathy-rubin-spectrum/</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=49999147</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49999147</guid></item><item><title><![CDATA[New comment by Reebz in "Fable 5.1 World Modeling"]]></title><description><![CDATA[
<p>being reductive, causality is the key term for a world model. the rollout is the key action to demonstrate causality. lots of nice LLMs can generate a 3d representation of a "visual world". These days the focus is on sensory data (image, video, etc.), however world models can be more than just 3d spatial representations - they could be tabular. Again being reductive, its how do you predict the next state from the current state (the rollout).</p>
]]></description><pubDate>Mon, 07 Sep 2026 07:17:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=49595003</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=49595003</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49595003</guid></item><item><title><![CDATA[New comment by Reebz in "Skimming an AI answer cost me 100 passwords"]]></title><description><![CDATA[
<p>My sense is yes. Running a small, non-frontier model in a "loose harness" has less guardrails under the hood. However I have no evidence besides my outcomes!<p>An analogy might be: AltaVisa 20+ years ago vs. Google + Chrome today. There are more layers to filter or warn of malicious links.</p>
]]></description><pubDate>Thu, 11 Jun 2026 00:04:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=48484533</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48484533</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48484533</guid></item><item><title><![CDATA[New comment by Reebz in "Skimming an AI answer cost me 100 passwords"]]></title><description><![CDATA[
<p>It was my fault, I copied the command and ran it from a search result provided by my OpenClaw that was running on a non-frontier model (I can’t remember which - something small and free).<p>I’m pretty confident it’s gone, this happened about 6-7 weeks ago. Have been running recurring checks for processes, Malwarebytes, and a PiHole to monitor traffic.</p>
]]></description><pubDate>Wed, 10 Jun 2026 22:12:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=48483437</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48483437</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48483437</guid></item><item><title><![CDATA[Skimming an AI answer cost me 100 passwords]]></title><description><![CDATA[
<p>Article URL: <a href="https://alwaysdraft.com/skimming-ai-answer-cost-100-passwords/">https://alwaysdraft.com/skimming-ai-answer-cost-100-passwords/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48477307">https://news.ycombinator.com/item?id=48477307</a></p>
<p>Points: 2</p>
<p># Comments: 4</p>
]]></description><pubDate>Wed, 10 Jun 2026 14:55:57 +0000</pubDate><link>https://alwaysdraft.com/skimming-ai-answer-cost-100-passwords/</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48477307</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48477307</guid></item><item><title><![CDATA[New comment by Reebz in "Claude Fable 5"]]></title><description><![CDATA[
<p>The Max version gets more details right. The bike frame looks good, the chain, the wings are appropriately styled instead of “arms”, and the knee is bent, etc. Obviously we’re hitting marginal returns now, but I see differences.</p>
]]></description><pubDate>Wed, 10 Jun 2026 03:52:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=48471218</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48471218</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48471218</guid></item><item><title><![CDATA[New comment by Reebz in "I design with Claude more than Figma now"]]></title><description><![CDATA[
<p>[Take a look at my portfolio site](<a href="https://reebz.com" rel="nofollow">https://reebz.com</a>), please view on desktop. This is about 3 weeks of effort to date. It is unfinished, but you get the idea.<p>Just like SaaS boilerplate from the decade prior, there is LLM boilerplate (since it’s trained on the internet).<p>So if you put in enough elbow-grease anything is (still) possible!</p>
]]></description><pubDate>Mon, 08 Jun 2026 09:08:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=48442948</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48442948</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48442948</guid></item><item><title><![CDATA[I don't believe in big decisions]]></title><description><![CDATA[
<p>Article URL: <a href="https://alwaysdraft.com/i-dont-believe-in-big-decisions/">https://alwaysdraft.com/i-dont-believe-in-big-decisions/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48357829">https://news.ycombinator.com/item?id=48357829</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Mon, 01 Jun 2026 15:06:17 +0000</pubDate><link>https://alwaysdraft.com/i-dont-believe-in-big-decisions/</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48357829</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48357829</guid></item><item><title><![CDATA[New comment by Reebz in "Show HN: My custom Statusline for Claude Code (Python wrapper around claudeline)"]]></title><description><![CDATA[
<p>There are certainly simpler solutions, but I love maximum flexibility to pickup and go from my desktop to iPad to iPhone anytime I want with full terminal access</p>
]]></description><pubDate>Thu, 21 May 2026 12:43:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=48221741</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48221741</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48221741</guid></item><item><title><![CDATA[New comment by Reebz in "Show HN: My custom Statusline for Claude Code (Python wrapper around claudeline)"]]></title><description><![CDATA[
<p>I use it regularly on my iPad with this setup: <a href="https://gist.github.com/Reebz/99db98ad4d3c45ebed84989a13710788" rel="nofollow">https://gist.github.com/Reebz/99db98ad4d3c45ebed84989a137107...</a></p>
]]></description><pubDate>Thu, 21 May 2026 04:01:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=48217717</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48217717</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48217717</guid></item><item><title><![CDATA[Show HN: My custom Statusline for Claude Code (Python wrapper around claudeline)]]></title><description><![CDATA[
<p>Hey everyone - After lots of tweaking and tinkering to find what I like, sharing my favorite way to see my Claude Code statusline. Hope you like it too.<p>The linked GH gist is a Python wrapper around claudeline that pipes its output through a few transformations and adds custom behaviors like a true-1M context %, and a custom context window alert  threshold with some emoji tweaks.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48216357">https://news.ycombinator.com/item?id=48216357</a></p>
<p>Points: 3</p>
<p># Comments: 4</p>
]]></description><pubDate>Thu, 21 May 2026 00:47:33 +0000</pubDate><link>https://gist.github.com/Reebz/741c5647c860fe0b5214f39d9d887240</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48216357</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48216357</guid></item><item><title><![CDATA[New comment by Reebz in "How Claude Code works in large codebases"]]></title><description><![CDATA[
<p>The influencer economy trades on hype, on frenzy, and ultimately, eyeballs. The more the better.<p>They want you feel like you’re missing out. They want you to switch. Being boring is far more productive. Pin your versions. Stick to stable releases and avoid the nightlies.<p>Significant noise created from 4.6 to 4.7 Opus transition has caused some to interpret this as signal. Excluding certain genuine and real bugs, the noise about perceived quality falling dramatically was noise. Influencers doing influencing turned it into “signal”. The reality was that if you had strong planning and spec driven development it ranged from manageable to non-existent.<p>The vast majority of the people I know and work with have not switched off CC or their Max sub.</p>
]]></description><pubDate>Fri, 15 May 2026 05:06:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=48144759</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=48144759</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48144759</guid></item><item><title><![CDATA[New comment by Reebz in "Grok 4.3"]]></title><description><![CDATA[
<p>Claude 4.7 is the clear winner to me for manager and formal report updates.<p>As an ex-senior exec (hundreds of staff), the bolded timeline impact is a particular nuance that I would expect a Lead/Director to format for a VP+ audience. Interesting none of the other models did that. My eyes immediately went to impact statement, then worked back to context to grasp the whole situation.</p>
]]></description><pubDate>Fri, 01 May 2026 23:53:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=47981861</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47981861</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47981861</guid></item><item><title><![CDATA[New comment by Reebz in "Show HN: My favorite local-feeling remotely accessible Claude Code setup"]]></title><description><![CDATA[
<p>Yeah fair question - I mainly did this because I’d forget to run /rc to enable remote control and the globally enabled rc flag is buggy.<p>The 3 main benefits for me are<p>- full terminal access, not just CC. So I can start CC remotely, not just join<p>- connection durability, CC sessions will die if network drops 10min+<p>- i enjoy cmux and wanted its workspace management integrated remotely as I’ve usually got 3-10 CC sessions active at any time</p>
]]></description><pubDate>Wed, 15 Apr 2026 21:14:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=47785317</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47785317</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47785317</guid></item><item><title><![CDATA[Show HN: My favorite local-feeling remotely accessible Claude Code setup]]></title><description><![CDATA[
<p>I run multiple Claude Code sessions on my Mac at home and control them from my iPhone or iPad while I’m out.<p>I like this because it gives me maximum power and flexibility (I can use any plugin or CLI as if I were sitting there locally).<p>The setup is cmux for the terminal multiplexer UI, tmux to keep sessions alive between connections, Tailscale for zero-config encrypted networking, and Echo app as the iOS SSH client. Optional Mosh for auto-reconnect when you switch networks.<p>The gist includes a ‘ccode’ alias shell function that handles session naming (so it doesn't conflict with claude if I ever need to use it), tmux lifecycle, continue/skip-permissions flags, and a pre-flight check so you don’t get a blank window when there’s nothing to continue.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47775039">https://news.ycombinator.com/item?id=47775039</a></p>
<p>Points: 4</p>
<p># Comments: 2</p>
]]></description><pubDate>Wed, 15 Apr 2026 05:31:28 +0000</pubDate><link>https://gist.github.com/Reebz/99db98ad4d3c45ebed84989a13710788</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47775039</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47775039</guid></item><item><title><![CDATA[Show HN: Signoff.sh – Claude Co-Authored-By with random fictional characters]]></title><description><![CDATA[
<p>Every Claude Code commit and PR is shipped with Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> (or similar). It's less fun than I think it should be. So signoff.sh is a 40-line bash script that replaces it with a random fictional character's full legal name - as best as I could figure out :)<p>Preserves the model name and the Anthropic noreply email, so you still get transparency about AI coding.<p>It's silly and that's the point.<p>I'm also a fan of peon ping (not mine): <a href="https://github.com/PeonPing/peon-ping" rel="nofollow">https://github.com/PeonPing/peon-ping</a></p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47766564">https://news.ycombinator.com/item?id=47766564</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Tue, 14 Apr 2026 15:02:05 +0000</pubDate><link>https://gist.github.com/Reebz/b3102c6a5de8238d3b60eb63450ee48e</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47766564</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47766564</guid></item><item><title><![CDATA[New comment by Reebz in "Ask HN: What are you building that's not AI related?"]]></title><description><![CDATA[
<p>I enjoyed this! Thanks for the index</p>
]]></description><pubDate>Thu, 09 Apr 2026 10:16:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=47701638</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47701638</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47701638</guid></item><item><title><![CDATA[New comment by Reebz in "MemPalace, the highest-scoring AI memory system ever benchmarked"]]></title><description><![CDATA[
<p>Many commentators are, mostly fairly, criticizing the repo due to issues raised on benchmarks. This is reasonable, however many are going further to bash the repo likely due to the authors.<p>For me, I see a silver lining. I'll be implementing mempalace for a few small agents to have memory portability that's managed locally.<p>I think the benchmarker who ran independent tests in GitHub issue #39 summed it up best:<p><i>To be clear about what this all means for our own use case: we still think there's a real product here, just not the one the README is selling. The combination of a one-command ChromaDB ingest pipeline for Claude Code, ChatGPT, and Slack exports, a working semantic search index over months or years of conversation history, fully local, MIT-licensed, no API key required, and a standalone temporal knowledge graph module (knowledge_graph.py) that could be used independently of the rest of the palace machinery,is genuinely useful, and we're planning to integrate it into our Sandcastle orchestrator as a claude_history_search MCP tool exactly along those lines.
</i><p><a href="https://github.com/milla-jovovich/mempalace/issues/39" rel="nofollow">https://github.com/milla-jovovich/mempalace/issues/39</a></p>
]]></description><pubDate>Tue, 07 Apr 2026 22:32:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=47682192</link><dc:creator>Reebz</dc:creator><comments>https://news.ycombinator.com/item?id=47682192</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47682192</guid></item></channel></rss>