<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: sorz</title><link>https://news.ycombinator.com/user?id=sorz</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 23 Apr 2026 01:28:39 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=sorz" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by sorz in "How does GPS work?"]]></title><description><![CDATA[
<p>So gound stations need keep measuring their own positions due to continental drift! Never thought that before. Thanks.</p>
]]></description><pubDate>Wed, 22 Apr 2026 13:02:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=47862971</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=47862971</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47862971</guid></item><item><title><![CDATA[New comment by sorz in "Lunar Flyby"]]></title><description><![CDATA[
<p>I found one taken with Z9: <a href="https://images.nasa.gov/details/art002e009301" rel="nofollow">https://images.nasa.gov/details/art002e009301</a><p>There are some very bright noise pixels on the dark area, which is different from the noise in similar photos taken with D5 (much darker and uniform).</p>
]]></description><pubDate>Wed, 08 Apr 2026 12:33:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=47689305</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=47689305</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47689305</guid></item><item><title><![CDATA[New comment by sorz in "What changes when you turn a Linux box into a router"]]></title><description><![CDATA[
<p>I feels wrong to not mention IPv6 in 2026.<p>- net.ipv6.conf.all.forwarding=1<p>- nftables is default to `ip` family which only applies to IPv4. Setting it to `inet` will allow rules to apply to both IPv4 & 6; or `ip6` for IPv6 only. You can skip NAT rules, usually.<p>- dnsmasq: in addition to DNS and DHCP, turns on router advertisement with SLAAC. Some devices can get IPv6 address from stateful DHCPv6 server, others (e.g. Android) only work with SLAAC.</p>
]]></description><pubDate>Sat, 04 Apr 2026 08:07:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=47636968</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=47636968</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47636968</guid></item><item><title><![CDATA[New comment by sorz in "Gpg.fail"]]></title><description><![CDATA[
<p>Lots of issues follow the pattern "ANSI escape code inside untrusted text". It feels like XSS but for terminal.</p>
]]></description><pubDate>Sun, 28 Dec 2025 02:36:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=46407856</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=46407856</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46407856</guid></item><item><title><![CDATA[New comment by sorz in "Tunnl.gg"]]></title><description><![CDATA[
<p>Random thoughts: one can get user's ssh public keys from GitHub on the fly (from `https://github.com/<username>.keys`), so that it requires a valid GitHub account to use this service, without (extra) auth process.</p>
]]></description><pubDate>Sat, 06 Dec 2025 07:40:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=46171448</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=46171448</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46171448</guid></item><item><title><![CDATA[Atomically accurate de novo design of antibodies with RFdiffusion]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.nature.com/articles/s41586-025-09721-5">https://www.nature.com/articles/s41586-025-09721-5</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=45829127">https://news.ycombinator.com/item?id=45829127</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 05 Nov 2025 22:48:07 +0000</pubDate><link>https://www.nature.com/articles/s41586-025-09721-5</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=45829127</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45829127</guid></item><item><title><![CDATA[New comment by sorz in "DOJ seizes $15B in Bitcoin from 'pig butchering' scam based in Cambodia"]]></title><description><![CDATA[
<p>Better be "three months of the country's GDP".</p>
]]></description><pubDate>Sat, 18 Oct 2025 11:33:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=45626518</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=45626518</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45626518</guid></item><item><title><![CDATA[New comment by sorz in "CAPTCHAs are over (in ticketing)"]]></title><description><![CDATA[
<p>The time at which scalpers buy tickets becomes irrelevant. They can resell tickets at arbitrary-high-price + margin because buyer know they will receive refund for the exceed amount, thus pay only (final price + scalper margin) for the ticket eventually.<p>Bonus: buy ticket from scalpers after price settled, you will get a determined price, no more guessing and find inner peace.<p>The problem for scalpers is that if they buy too many tickets, the final price may become too high to be attractive for real buyers.</p>
]]></description><pubDate>Mon, 26 May 2025 14:32:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=44097795</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=44097795</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44097795</guid></item><item><title><![CDATA[New comment by sorz in "Magma: A foundation model for multimodal AI agents"]]></title><description><![CDATA[
<p>In the mug-scrubbing video, the person clearly <i>pretends</i> to wash the cup but does not seem to want to get their hands wet anyway. I'm curious as to when models can figure out that subtle thing.</p>
]]></description><pubDate>Thu, 20 Feb 2025 08:48:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=43112514</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=43112514</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43112514</guid></item><item><title><![CDATA[New comment by sorz in "Learn AutoHotKey by stealing my scripts"]]></title><description><![CDATA[
<p>Did you try scancode map registry[1]? I use it to swap Ctrl/CapsLock and it works well even with RDP softwares and run-as-admin apps which PowerToys couldn't handle.<p>[1] <a href="https://superuser.com/questions/550679/where-can-i-find-windows-keyboard-scancode-registry-information" rel="nofollow noreferrer">https://superuser.com/questions/550679/where-can-i-find-wind...</a></p>
]]></description><pubDate>Tue, 22 Aug 2023 05:29:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=37219025</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=37219025</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=37219025</guid></item><item><title><![CDATA[New comment by sorz in "Has Google Translate been fixed yet?"]]></title><description><![CDATA[
<p>The whole post is on base on the fact that "TW" in their URL. But I think this is really just a internal implementation detail of Google Translate, which is not intended to display to users (e.g. complete hidden on their app). Everywhere on their UI shows "Chinese (Traditional)". That is what they try to communicate with users.<p>Sure, zh-TW is somewhat misleading. But they nor say that parameter is a ISO 639 or RFC 5646 conformed.</p>
]]></description><pubDate>Fri, 11 Aug 2023 04:40:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=37085172</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=37085172</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=37085172</guid></item><item><title><![CDATA[New comment by sorz in "RWKV: Reinventing RNNs for the Transformer Era"]]></title><description><![CDATA[
<p>They said the paper is still working in progress and will improve it.<p><a href="https://twitter.com/AiEleuther/status/1660811180901019648" rel="nofollow">https://twitter.com/AiEleuther/status/1660811180901019648</a></p>
]]></description><pubDate>Tue, 23 May 2023 07:54:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=36041280</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=36041280</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=36041280</guid></item><item><title><![CDATA[New comment by sorz in "ChatGPT outperforms crowd-workers for text-annotation tasks"]]></title><description><![CDATA[
<p>Just watched a talk[0] about natural language understanding research in post-GPT-3 era. Old issues may has been solved, while new topics are coming to this area (quoted from the slides):<p>- Retrieval augmented in-context learning<p>- Better benchmarks<p>- Last mile for productive application<p>- Faithful, human-interoperable explanations<p>[0] <a href="https://www.youtube.com/watch?v=-lnHHWRCDGk">https://www.youtube.com/watch?v=-lnHHWRCDGk</a></p>
]]></description><pubDate>Thu, 30 Mar 2023 09:12:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=35369986</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=35369986</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35369986</guid></item><item><title><![CDATA[New comment by sorz in "ChatGPT outperforms crowd-workers for text-annotation tasks"]]></title><description><![CDATA[
<p>Is NLP a solved problem now?</p>
]]></description><pubDate>Tue, 28 Mar 2023 02:55:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=35335300</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=35335300</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35335300</guid></item><item><title><![CDATA[New comment by sorz in "BlenderGPT: Use commands in English to control Blender with OpenAI's GPT-4"]]></title><description><![CDATA[
<p>> then the amount of information we need to make billboards and signs make sense.<p>Subsequently, this applies to posters, letters, newspapers, and other types of text-heavy images, ultimately reducing the language modeling problem to an image generation problem.</p>
]]></description><pubDate>Mon, 27 Mar 2023 07:34:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=35323912</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=35323912</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35323912</guid></item><item><title><![CDATA[New comment by sorz in "GPTs Are GPTs: An Early Look at the Labor Market Impact Potential of LLMs"]]></title><description><![CDATA[
<p>OpenAI did exact the test for GPT-4. The raw, non-fine-tune GPT-4 is quite good at predicting confidence level ("highly calibrated" by their words). But the RLHF fine-tuning process seems ruin its calibration. Figure 8 on page 12 of GPT-4 Technical Report shows this dramatic changes before & after fine-tuning.</p>
]]></description><pubDate>Mon, 20 Mar 2023 16:52:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=35234532</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=35234532</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35234532</guid></item><item><title><![CDATA[New comment by sorz in "How Rust 1.64 became faster on Windows"]]></title><description><![CDATA[
<p>Is it possible that the profile is over-fitting to the benchmark tests?</p>
]]></description><pubDate>Sun, 23 Oct 2022 16:23:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=33308102</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=33308102</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=33308102</guid></item><item><title><![CDATA[New comment by sorz in "Guest WiFi using a QR code"]]></title><description><![CDATA[
<p>Does it work for WPA3-Personal network?</p>
]]></description><pubDate>Wed, 13 Jul 2022 09:34:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=32080544</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=32080544</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=32080544</guid></item><item><title><![CDATA[New comment by sorz in "Show HN: Flash your ESP32 from the browser using JavaScript"]]></title><description><![CDATA[
<p>Want to know when Firefox will support Web Serial API, but turn out they won't since they consider it harmful: <a href="https://mozilla.github.io/standards-positions/#webserial" rel="nofollow">https://mozilla.github.io/standards-positions/#webserial</a></p>
]]></description><pubDate>Wed, 07 Jul 2021 09:56:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=27759293</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=27759293</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=27759293</guid></item><item><title><![CDATA[One file that mixes all current JavaScript proposals]]></title><description><![CDATA[
<p>Article URL: <a href="https://twitter.com/jackworks_asref/status/1330172789924442112">https://twitter.com/jackworks_asref/status/1330172789924442112</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=25173916">https://news.ycombinator.com/item?id=25173916</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Sun, 22 Nov 2020 00:56:33 +0000</pubDate><link>https://twitter.com/jackworks_asref/status/1330172789924442112</link><dc:creator>sorz</dc:creator><comments>https://news.ycombinator.com/item?id=25173916</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=25173916</guid></item></channel></rss>