<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: lukah</title><link>https://news.ycombinator.com/user?id=lukah</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Mon, 04 May 2026 16:09:36 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=lukah" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by lukah in "OpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors"]]></title><description><![CDATA[
<p>Exactly this. Most diagnosis isn’t about pinpointing the underlying exact cause, it’s ruling out the really bad stuff and minimising harm. Differential diagnosis just isn’t real world medicine.</p>
]]></description><pubDate>Mon, 04 May 2026 09:37:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48006436</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=48006436</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48006436</guid></item><item><title><![CDATA[New comment by lukah in "Jury told that Meta, Google 'engineered addiction' at landmark US trial"]]></title><description><![CDATA[
<p>I'm not at all familiar with the American justice system, but would the fact that this case specifically describes the targeting of minors with such addictive tactics change things at all?</p>
]]></description><pubDate>Tue, 10 Feb 2026 15:53:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=46961505</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=46961505</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46961505</guid></item><item><title><![CDATA[New comment by lukah in "Anthropic's original take home assignment open sourced"]]></title><description><![CDATA[
<p>I've never heard AI-assisted coding referred to as "RalphWiggum"ing a problem, and now I will have to use that always. Thank you.</p>
]]></description><pubDate>Wed, 21 Jan 2026 12:54:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=46705054</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=46705054</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46705054</guid></item><item><title><![CDATA[New comment by lukah in "Why proteins fold and how GPUs help us fold"]]></title><description><![CDATA[
<p>I interpreted that section as alphafold not learning physics, but rather correlations within a constrained setting that a-priori correspond to physically sound inferences. It has a specific architecture that allows the model to make inferences that are more physically plausible than not, but not that it’s discovering actual, causally verifiable laws of nature (like what I’d assume are encoded into another non-ML approach to the folding problem for example).</p>
]]></description><pubDate>Mon, 15 Dec 2025 10:31:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=46272664</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=46272664</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46272664</guid></item><item><title><![CDATA[New comment by lukah in "Transformers Without Normalization"]]></title><description><![CDATA[
<p>From their implementation it looks like they’re calculating tanh and then applying a weight and bias</p>
]]></description><pubDate>Sat, 15 Mar 2025 06:52:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=43370567</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=43370567</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43370567</guid></item><item><title><![CDATA[New comment by lukah in "Ocean Carbon Removal: Captura's marine carbon capture explained"]]></title><description><![CDATA[
<p>It’s true that this could lead to unintended consequences as humanity is directly experimenting with a complex balanced natural system. But another perspective is that we have already done/are doing exactly that by pumping CO2 into the atmosphere at rates that wouldn’t have happened in nature. I know “two wrongs don’t make a right”, but it’s perhaps worth considering the price of course correcting our already hugely impactful behaviours, as the butterfly had flown years ago.</p>
]]></description><pubDate>Fri, 27 Dec 2024 11:49:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=42521467</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=42521467</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42521467</guid></item><item><title><![CDATA[New comment by lukah in "Cybertruck's Many Recalls"]]></title><description><![CDATA[
<p>Car panels are convex on nearly all other cars for very good reason. Flat panels are structurally susceptible to damages which wouldn’t mark a standard panel. Adding a highly reflective surface was another great move.</p>
]]></description><pubDate>Tue, 26 Nov 2024 06:49:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=42243088</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=42243088</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42243088</guid></item><item><title><![CDATA[New comment by lukah in "The Power of Small Brain Networks"]]></title><description><![CDATA[
<p>How do they track the brain activity of a fruit fly? That seems incredible to me</p>
]]></description><pubDate>Mon, 25 Nov 2024 06:24:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=42233750</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=42233750</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42233750</guid></item><item><title><![CDATA[New comment by lukah in "Show HN: Using GPT-3 and Whisper to save doctors’ time"]]></title><description><![CDATA[
<p>This is really interesting work, congratulations! A lot of negativity in the comments seems to stem from cynicism around doctors and their billing, but in a country like the UK I know removing essentially "admin tasks" from the GP would be very welcome and give time back to the clinician, allowing them to spend more time focussing on the patient and outcomes.</p>
]]></description><pubDate>Tue, 14 Mar 2023 15:08:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=35152706</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=35152706</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35152706</guid></item><item><title><![CDATA[New comment by lukah in "Understanding and coding the self-attention mechanism of large language models"]]></title><description><![CDATA[
<p>Great answer. Imbuing a deep learning model with well thought out inductive biases is one of the strongest ways of guiding your model to interpret the data the way you want it to. Otherwise it’s kind of shooting in the dark and hoping to get lucky.<p>I can’t stand it when people lazily personify ML models, but it’s akin to giving someone with no experience  some wood and then pointing to a shed and saying “make one of those from this”. Instead you’d expect them to be much more successful if you also give them a saw, a drill, some screws etc.</p>
]]></description><pubDate>Sat, 11 Feb 2023 07:25:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=34750337</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=34750337</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34750337</guid></item><item><title><![CDATA[New comment by lukah in "Aerodynamics of Gravel Bikes"]]></title><description><![CDATA[
<p>If anything, you might be thankful of a full suspension mountain bike on some of our "gravel" trails in the UK. There's such a wide range from US-style back county roads all the way to bridleways with baby-head sized rocks that you probably wouldn't want to be riding on something typically sold as a gravel bike.</p>
]]></description><pubDate>Tue, 20 Sep 2022 12:36:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=32910814</link><dc:creator>lukah</dc:creator><comments>https://news.ycombinator.com/item?id=32910814</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=32910814</guid></item></channel></rss>