<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: snemvalts</title><link>https://news.ycombinator.com/user?id=snemvalts</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 02 Jul 2026 16:18:09 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=snemvalts" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by snemvalts in "ZCode – Harness for GLM-5.2"]]></title><description><![CDATA[
<p>But this harness app is chinese?</p>
]]></description><pubDate>Thu, 02 Jul 2026 05:29:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=48756902</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48756902</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48756902</guid></item><item><title><![CDATA[New comment by snemvalts in "LLMs are eroding my software engineering career and I don't know what to do"]]></title><description><![CDATA[
<p>Most benchmarks can be trained for as well, so they are over-representative of model's engineering skills. 
The entire nature of a benchmark is collapsing some qualitative work (software engineering task, architecture choice, code quality) into a quantitative score which can be optimized for.</p>
]]></description><pubDate>Mon, 08 Jun 2026 07:51:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=48442451</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48442451</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48442451</guid></item><item><title><![CDATA[New comment by snemvalts in "Does anybody like React?"]]></title><description><![CDATA[
<p>Let's say you to render conditionally when `x` is present. You do `v-if="x"`. Now you want to refer to `x.y.z` in the body, while `y` is optional.<p>Can you be sure that the value is present? How do you check for it? What does Vue need to do to enable this functionality with static type checking with TypeScript?<p>With `{x && <>{x.y.z}</>}`, it's almost vanilla TS, it's the same as `x && x.y.z` in normal TS. Type narrowings that work in TS are guaranteed to work in React, without the framework having to do anything.<p>Less framework/library code required to make the same thing work. To me, that's the sign of a better abstraction and implementation.</p>
]]></description><pubDate>Mon, 01 Jun 2026 08:24:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=48354028</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48354028</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48354028</guid></item><item><title><![CDATA[New comment by snemvalts in "Does anybody like React?"]]></title><description><![CDATA[
<p>JSX is a very thin layer of templating logic inside JS. Meaning you have all the language features available while templating. Some arbitrary JS can result in templating.<p>Vue's DSL is whatever language the developer implemented. Which is probably not enough, depends how much effort they put into it and how good they are in language design. Given that they cargo cult HTML tags to organize components in a pseudo-familiar but not-valid-HTML way, I don't have much confidence in their language design skills.<p>I'd take the former any day.</p>
]]></description><pubDate>Tue, 26 May 2026 06:59:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=48276087</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48276087</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48276087</guid></item><item><title><![CDATA[New comment by snemvalts in "Jira Is Turing-Complete"]]></title><description><![CDATA[
<p>How many of them can run infinitely? Or be re-triggered by humans to continue where they left off?</p>
]]></description><pubDate>Mon, 25 May 2026 06:01:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=48263864</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48263864</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48263864</guid></item><item><title><![CDATA[New comment by snemvalts in "An OpenAI model has disproved a central conjecture in discrete geometry"]]></title><description><![CDATA[
<p>Tai's Model (<a href="https://en.wikipedia.org/wiki/Tai%27s_model" rel="nofollow">https://en.wikipedia.org/wiki/Tai%27s_model</a>) is a perfect example of this</p>
]]></description><pubDate>Thu, 21 May 2026 05:00:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=48218065</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48218065</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48218065</guid></item><item><title><![CDATA[New comment by snemvalts in "Access to frontier AI will soon be limited by economic and security constraints"]]></title><description><![CDATA[
<p>What about other benchmarks? Benchmarks where the contents are freely available  have become useless for evaluating models.</p>
]]></description><pubDate>Fri, 15 May 2026 14:43:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=48149302</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=48149302</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48149302</guid></item><item><title><![CDATA[New comment by snemvalts in "Epoch confirms GPT5.4 Pro solved a frontier math open problem"]]></title><description><![CDATA[
<p>Scaling law is a power law , requiring orders of magnitude more compute and data for better accuracy from pre-training. Most companies have maxed it out.<p>For RL, we are arriving at a similar point
<a href="https://www.tobyord.com/writing/how-well-does-rl-scale" rel="nofollow">https://www.tobyord.com/writing/how-well-does-rl-scale</a><p>Next stop is inference scaling  with longer context window and longer reasoning. But instead of it being a one-off training cost, it becomes a running cost.<p>In essence we are chasing ever smaller gains in exchange for exponentially increasing costs. This energy will run out. There needs to be something completely different than LLMs for meaningful further progress.</p>
]]></description><pubDate>Tue, 24 Mar 2026 08:54:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=47500059</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=47500059</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47500059</guid></item><item><title><![CDATA[New comment by snemvalts in "Epoch confirms GPT5.4 Pro solved a frontier math open problem"]]></title><description><![CDATA[
<p>The ability to learn and infer without absorbing millions of books and all text on internet really does make us special. And only at 20 watts!</p>
]]></description><pubDate>Tue, 24 Mar 2026 05:09:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=47498807</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=47498807</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47498807</guid></item><item><title><![CDATA[New comment by snemvalts in "Epoch confirms GPT5.4 Pro solved a frontier math open problem"]]></title><description><![CDATA[
<p>Math and coding competition problems are easier to train because of strict rules and cheap verification.  
But once you go beyond that to less defined things such as code quality, where even humans have hard time putting down concrete axioms, they start to hallucinate more and become less useful.<p>We are missing the value function that allowed AlphaGo to go from mid range player trained on human moves to superhuman by playing itself.  
As we have only made progress on unsupervised learning, and RL is constrained as above, I don't see this getting better.</p>
]]></description><pubDate>Tue, 24 Mar 2026 04:59:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=47498755</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=47498755</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47498755</guid></item><item><title><![CDATA[New comment by snemvalts in "Ex-GitHub CEO launches a new developer platform for AI agents"]]></title><description><![CDATA[
<p>In simpler terms - they create an MCP server, essentially an API that the coding agent can call, that can fill in context about previous decisions done by the coding agent earlier in development. Agent equivalent of asking someone who's been working there longer "why is this this way".<p>This means that the agent will can have context of previous decisions, something that they currently struggle with as they are always starting from a blank slate.</p>
]]></description><pubDate>Sat, 14 Feb 2026 10:55:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=47013489</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=47013489</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47013489</guid></item><item><title><![CDATA[New comment by snemvalts in "Ex-GitHub CEO launches a new developer platform for AI agents"]]></title><description><![CDATA[
<p>MCP server with RAG to feed it back to agents when they are working on a piece of code, and bob's your uncle</p>
]]></description><pubDate>Wed, 11 Feb 2026 06:57:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=46971760</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=46971760</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46971760</guid></item><item><title><![CDATA[New comment by snemvalts in "Sugar industry influenced researchers and blamed fat for CVD (2016)"]]></title><description><![CDATA[
<p>Treating T2D and preventing T2D are completely different things from a dietary perspective. Same way you wouldn't give chemotherapy to a healthy person to prevent cancer</p>
]]></description><pubDate>Thu, 08 Jan 2026 10:54:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=46539588</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=46539588</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46539588</guid></item><item><title><![CDATA[New comment by snemvalts in "Sugar industry influenced researchers and blamed fat for CVD (2016)"]]></title><description><![CDATA[
<p>Red meat (a known carcinogen) at the top is gold. All that saturated fat the energy will come from (not from protein or veggies) will probably cause heart problems and plaque formation in arteries, not to mention insulin resistance <i>just</i> from increased FFAs in blood.<p>Vegetarians and vegans have lower T2D incidence on average FWIW.</p>
]]></description><pubDate>Wed, 07 Jan 2026 20:37:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=46532276</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=46532276</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46532276</guid></item><item><title><![CDATA[New comment by snemvalts in "Show HN: Autism Simulator"]]></title><description><![CDATA[
<p>i would interpret physical fitness as cardio exercise routine and depleted muscle glycogen stores:
 so breakfast is very welcome and without it is not possible to keep up exercise routine</p>
]]></description><pubDate>Thu, 02 Oct 2025 05:43:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=45446651</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=45446651</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45446651</guid></item><item><title><![CDATA[New comment by snemvalts in "React is winning by default and slowing innovation"]]></title><description><![CDATA[
<p>Yes, but it is not syntax. It's a contract with the library. React is completely usable using vanilla JS syntax. Same cannot be said for Vue and Angular.
It feels a bit like talking about apples and oranges in this thread.</p>
]]></description><pubDate>Wed, 17 Sep 2025 06:15:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=45272369</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=45272369</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45272369</guid></item><item><title><![CDATA[New comment by snemvalts in "React is winning by default and slowing innovation"]]></title><description><![CDATA[
<p>You can write const [a, b] = useState('x') in vanilla js and typescript. Hence it is not magic <i>syntax</i>.</p>
]]></description><pubDate>Tue, 16 Sep 2025 14:47:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=45263033</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=45263033</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45263033</guid></item><item><title><![CDATA[New comment by snemvalts in "Apple's On-Device and Server Foundation Models"]]></title><description><![CDATA[
<p>The swapping is indeed faster as the SSD is on the SoC and so fast to access.
To the point that an 4 year old 8gb M1 Air is enough for simpler development work, at least for me.</p>
]]></description><pubDate>Tue, 11 Jun 2024 03:49:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=40642102</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=40642102</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40642102</guid></item><item><title><![CDATA[New comment by snemvalts in "Cardio fitness is a strong, consistent predictor of morbidity and mortality"]]></title><description><![CDATA[
<p>I'm speaking from experience regarding mental illnesses and exercise. And I never discounted medication. 
Just that exercise is critically underprescribed, I'm fairly sure it would work better for milder cases compared to meds. Not to mention the other health benefits listed in the thread.<p>Same way an opioid pill is still prescribed in cases of cancer or severe pain. Just that there are probably better, milder alternatives that don't have as many side effects that could fit a lot of these people with milder problems.</p>
]]></description><pubDate>Fri, 03 May 2024 08:10:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=40245360</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=40245360</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40245360</guid></item><item><title><![CDATA[New comment by snemvalts in "Cardio fitness is a strong, consistent predictor of morbidity and mortality"]]></title><description><![CDATA[
<p>If you have to measure time spent and do cost/benefit analysis, you're most likely not going to stick with it for the entirety of life.</p>
]]></description><pubDate>Fri, 03 May 2024 04:50:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=40244163</link><dc:creator>snemvalts</dc:creator><comments>https://news.ycombinator.com/item?id=40244163</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40244163</guid></item></channel></rss>