<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: vgalin</title><link>https://news.ycombinator.com/user?id=vgalin</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 09 Sep 2026 15:03:41 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=vgalin" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by vgalin in "ChatGPT Images 2.5"]]></title><description><![CDATA[
<p>The introductory video is a bit... daring. Generating tattoos that clearly look AI-generated is not one of the use-cases I'd try to sell.</p>
]]></description><pubDate>Tue, 08 Sep 2026 19:13:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=49615317</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=49615317</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49615317</guid></item><item><title><![CDATA[New comment by vgalin in "Google says criminal hackers used AI to find a major software flaw"]]></title><description><![CDATA[
<p>The company doing the actual ID verification (KYC) is probably the last company I'd trust with this kind of data.<p>To circumvent conversations being flagged as "cybersecurity bad!!!" I often have to use previous models (5.3 for example, and sometimes using them through subagents is enough). And when this method no longer works, local models will be good enough for it to not be a problem (for my use case, at least).</p>
]]></description><pubDate>Tue, 12 May 2026 00:46:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=48102778</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=48102778</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48102778</guid></item><item><title><![CDATA[New comment by vgalin in "Google releases Gemma 4 open models"]]></title><description><![CDATA[
<p>I ran gemma4:26b without any tooling access and it gave me the correct answer in a few minutes only (definitely less than 8 minutes, but I didn't timed it).<p>Specs : RX 9070 XT (24GB VRAM) + 16 GB RAM<p>gist : <a href="https://gist.github.com/vgalin/a9c852605f39ab503f167c9708a46466" rel="nofollow">https://gist.github.com/vgalin/a9c852605f39ab503f167c9708a46...</a><p>(I gave it another go and it found the correct result in about a minute, see the comment on the gist)</p>
]]></description><pubDate>Thu, 02 Apr 2026 22:50:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=47621203</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=47621203</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47621203</guid></item><item><title><![CDATA[New comment by vgalin in "Switch to Claude without starting over"]]></title><description><![CDATA[
<p>> And that is shared with chat too.<p>Huh, I didn't know about that. I'm trying Claude Pro for the first time while comparing it against ChatGPT and I'm (sadly) not impressed at the moment.<p>When I asked both Codex and Claude Code to "look into" an issue of medium-to-high-complexity in a code base, Codex went with the fix I had in mind and directly and made code changes without being asked or at least asking for permission. It only used a few percents of its 5-hour limits to do it, on `High`.<p>Claude in the meanwhile misdiagnosed the core of the issue on its first pass (even on Opus 4.6 + Thinking). I had to guide it in the right direction and despite being given the 'answer', it was quite a long process compared to Codex' one-shot. And it hit the 5h limit before being able to finish solving the issue.</p>
]]></description><pubDate>Sun, 01 Mar 2026 19:10:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=47209680</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=47209680</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47209680</guid></item><item><title><![CDATA[New comment by vgalin in "Pyxel: A retro game engine for Python"]]></title><description><![CDATA[
<p>Python library with Rust backend*</p>
]]></description><pubDate>Thu, 11 Jul 2024 07:55:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=40934542</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=40934542</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40934542</guid></item><item><title><![CDATA[New comment by vgalin in "I am sick of LeetCode-style interviews"]]></title><description><![CDATA[
<p>But the older you are, the more likely it is that you have built a family and have children and responsibilities outside of work. You cannot necessarily afford to spend time studying (problems untied to the real world) as much as someone younger I guess.</p>
]]></description><pubDate>Tue, 04 Jun 2024 12:02:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=40573548</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=40573548</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40573548</guid></item><item><title><![CDATA[New comment by vgalin in "Making USB devices – end to end guide to your first gadget"]]></title><description><![CDATA[
<p>I didn't used a kit, but bought the components separately (microcontroller, key caps, switches, wires). It runs custom code as it was sufficient to fit my use case, though QMK may be possible? The keys are mapped from F13 to F22 and can be use in some software that allow custom keybinds.</p>
]]></description><pubDate>Tue, 04 Jun 2024 07:05:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=40571480</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=40571480</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40571480</guid></item><item><title><![CDATA[New comment by vgalin in "Making USB devices – end to end guide to your first gadget"]]></title><description><![CDATA[
<p>I recently built a small hand-wired macro pad using an Arduino Pro Micro equipped with ATmega32U4, it's apparently quite popular amongst hobbyists building custom keyboards. Quick and fun project for a beginner, the most tedious part of this project was to carve the wooden case.</p>
]]></description><pubDate>Mon, 03 Jun 2024 11:34:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=40561648</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=40561648</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40561648</guid></item><item><title><![CDATA[New comment by vgalin in "Colab notebook to create Magic cards from image with Claude"]]></title><description><![CDATA[
<p>(Disclaimer, I'm the maintainer of this package, but this kind of use is exactly why I created it at the beginning)<p>If you know how to use HTML+CSS and would like to generate full-fledged cards, you could use a package such as html2image [0] to combine the text, the image and a card-template image into one final image.
Chrome/Chromium has to be available on Colab Notebooks though, that's the only requirement.
Using basic SVG without this package could also do the trick.<p>[0] <a href="https://github.com/vgalin/html2image">https://github.com/vgalin/html2image</a></p>
]]></description><pubDate>Mon, 08 Apr 2024 21:36:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=39973987</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=39973987</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=39973987</guid></item><item><title><![CDATA[New comment by vgalin in "Algorithms for Modern Hardware"]]></title><description><![CDATA[
<p>People are now boycotting knowledge?</p>
]]></description><pubDate>Sat, 16 Mar 2024 14:24:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=39726207</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=39726207</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=39726207</guid></item><item><title><![CDATA[New comment by vgalin in "Thanks FedEx, this is why we keep getting phished"]]></title><description><![CDATA[
<p>(translation provided by ChatGPT)<p>> Terms and Conditions, Price and Service List, Conditions.<p>> Dear customer,<p>> our price and service list, our terms and conditions, as well as further conditions which will come into effect on May 1, 2024, can be found on the USB stick.<p>> With kind regards,<p>> The Sparkasse Bremen AG</p>
]]></description><pubDate>Fri, 23 Feb 2024 12:10:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=39479589</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=39479589</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=39479589</guid></item><item><title><![CDATA[New comment by vgalin in "Ask HN: What's your "it's not stupid if it works" story?"]]></title><description><![CDATA[
<p>I wrote a Python package [1] that does something similar! It allows the generation of images from HTML+CSS strings or files (or even other files like SVGs) and could probably handle PDF generation too.
It uses the headless version of Chrome/Chromium or Edge behind the scenes.<p>Writing this package made me realize that even big projects (such as Chromium) sometimes have features that just <i>don't</i> work. Edge headless wouldn't let you take screenshots up until recently, and I still encountered issues with Firefox last time I tried to add support for it in the package. I also stumbled upon weird behaviors of Chrome CDP when trying to implement an alternative to using the headless mode, and these issues eventually fixed themselves after some Chrome updates.<p>[1] <a href="https://github.com/vgalin/html2image">https://github.com/vgalin/html2image</a></p>
]]></description><pubDate>Fri, 22 Dec 2023 13:41:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=38734022</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=38734022</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=38734022</guid></item><item><title><![CDATA[New comment by vgalin in "How would you say “She said goodbye too many times before.” in Latin?"]]></title><description><![CDATA[
<p>For further nitpicking: "etc." shall also end with '.' as it is the abbreviated form of "et cetera/caetera". Also, when used in an enumeration, it shall be preceded by a comma.<p><a href="https://en.wikipedia.org/wiki/Et_cetera" rel="nofollow noreferrer">https://en.wikipedia.org/wiki/Et_cetera</a></p>
]]></description><pubDate>Wed, 06 Sep 2023 12:29:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=37404107</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=37404107</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=37404107</guid></item><item><title><![CDATA[New comment by vgalin in "Harvard ethics professor allegedly fabricated multiple studies"]]></title><description><![CDATA[
<p>If this is an established fact, where could one learn more about this?</p>
]]></description><pubDate>Mon, 10 Jul 2023 14:05:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=36666646</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=36666646</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=36666646</guid></item><item><title><![CDATA[New comment by vgalin in "Bark: A transformer based text to audio system"]]></title><description><![CDATA[
<p>The french example sounds off in my opinion, it seems to mix up its consonants/ vowels.<p>The "OU" sound from "nous" sounds like "EU", or like if the sound was skipped.  
The "OH" sound from "sommes" sounds like "EU" (again, but less pronounced).  
The "T" sound from "trop" becames "P" ("pro").</p>
]]></description><pubDate>Mon, 15 May 2023 09:50:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=35945759</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=35945759</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35945759</guid></item><item><title><![CDATA[New comment by vgalin in "Ask HN: Where to Feature Startup on Spanish, Portuguese, German, French?"]]></title><description><![CDATA[
<p>It appears to be invite-only (even though you can request an invite). I'm not sure about what drove them to do so, but that may be why there is not much activity on the website.</p>
]]></description><pubDate>Fri, 12 May 2023 13:54:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=35916351</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=35916351</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35916351</guid></item><item><title><![CDATA[New comment by vgalin in "On Having Enough Socks (2019)"]]></title><description><![CDATA[
<p>Interestingly, they have a "barcode" system that allow them (and allow us) to know the age of a given sock.<p>"Blacksocks knows the age of all socks"
<a href="https://www.blacksocks.com/en/ageofthesocksen" rel="nofollow">https://www.blacksocks.com/en/ageofthesocksen</a></p>
]]></description><pubDate>Fri, 05 May 2023 13:40:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=35829060</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=35829060</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35829060</guid></item><item><title><![CDATA[New comment by vgalin in "Show HN: GPT-4-powered web searches for developers"]]></title><description><![CDATA[
<p>This peculiar tool seems to be aimed at developers, so I don't find it surprising.</p>
]]></description><pubDate>Thu, 13 Apr 2023 08:31:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=35552583</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=35552583</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35552583</guid></item><item><title><![CDATA[New comment by vgalin in "Attacker with access to XML config can trigger keepass.exe to obtain passwords"]]></title><description><![CDATA[
<p>If 'someone' (an attacker), introduced themselves physically in your house or workplace (thus obtaining a high level of access) to place hidden cameras and to record you typing your master passwords. And if that someone later retrieved the footage and a copy of your password manager database... Would you blame your password manager for being insecure?</p>
]]></description><pubDate>Fri, 27 Jan 2023 13:51:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=34545986</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=34545986</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34545986</guid></item><item><title><![CDATA[New comment by vgalin in "Attacker with access to XML config can trigger keepass.exe to obtain passwords"]]></title><description><![CDATA[
<p>Then it would be an attack vector of course, but whether it is a vulnerability of the password manager itself would be debatable.</p>
]]></description><pubDate>Fri, 27 Jan 2023 13:33:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=34545766</link><dc:creator>vgalin</dc:creator><comments>https://news.ycombinator.com/item?id=34545766</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34545766</guid></item></channel></rss>