<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: loehnsberg</title><link>https://news.ycombinator.com/user?id=loehnsberg</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Tue, 14 Apr 2026 10:07:56 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=loehnsberg" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by loehnsberg in "OpenClaw’s memory is unreliable, and you don’t know when it will break"]]></title><description><![CDATA[
<p>I think if we want to build on what we have, instead of compaction at the end of the context window, the LLM would have to 'sleep', i.e. adjust its weights, then wake up with the last bits of the old context window in the new one, and have a 'feel' for what it did before through the change in weights. I just sense it's not that simple to get there, because simply updating the weights based on a single context sample risks degrading the weights of the whole network.<p>I like the idea of using small local model (or several) for tackling this problem, like low rank adaptation, but with current tech, I still have to piece this together or the small local models will forget old memories.</p>
]]></description><pubDate>Sat, 11 Apr 2026 09:17:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=47728946</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=47728946</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47728946</guid></item><item><title><![CDATA[New comment by loehnsberg in "OpenClaw’s memory is unreliable, and you don’t know when it will break"]]></title><description><![CDATA[
<p>As long as there's no solution to the long-term memory problem, we will have a "country of geniuses in a data center" that are all suffering from anterograde amnesia (movie: Memento), which requires human hand-holding.<p>I have experimented with a lot of hacks, like hierarchies of indexed md files, semantic DBs, embeddings, dynamic context retrieval, but none of this is really a comprehensive solution to get something that feels as intelligent as what these systems are able to do within their context windows.<p>I am als a touch skeptical that adjusting weights to learn context will do the trick without a transformer-like innovation in reinforcement learning.<p>Anyway, I‘ll keep tinkering…</p>
]]></description><pubDate>Sat, 11 Apr 2026 07:50:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47728437</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=47728437</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47728437</guid></item><item><title><![CDATA[New comment by loehnsberg in "European alternatives to Google, Apple, Dropbox and 120 US apps"]]></title><description><![CDATA[
<p>I guess you can make that same argument about USA and America. Canada is clearly America but a Canadian would not refer to himself as American whereas a US national would. Europeans hardly refer to themselves as such but when European countries are lumped together, it has become common to ignore geography and refer to those affiliated with EU membership or bilateral EU affiliation (Norway, Switzerland, Iceland) as European.</p>
]]></description><pubDate>Fri, 03 Apr 2026 11:20:34 +0000</pubDate><link>https://news.ycombinator.com/item?id=47625386</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=47625386</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47625386</guid></item><item><title><![CDATA[New comment by loehnsberg in "My Random Forest Was Mostly Learning Time-to-Expiry Noise"]]></title><description><![CDATA[
<p>The blogger trained a random forest on the returns of some price with a set of features and the ttm feature was strongest. Did I miss anything?</p>
]]></description><pubDate>Fri, 20 Mar 2026 04:57:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=47450670</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=47450670</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47450670</guid></item><item><title><![CDATA[New comment by loehnsberg in "Hegseth gives Anthropic until Friday to back down on AI safeguards"]]></title><description><![CDATA[
<p>I really hope they continue to show some spine against this administration and do not allow to weaponize AI against human beings. It's the morally right thing to do!</p>
]]></description><pubDate>Tue, 24 Feb 2026 22:39:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=47144352</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=47144352</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47144352</guid></item><item><title><![CDATA[New comment by loehnsberg in "European Alternatives"]]></title><description><![CDATA[
<p>Isn't it sad that we now have Russian, Chinese, American, European, etc alternatives? I mean I get it, Sept 11 paved the way for FISA orders and NSA overreach, Russia and China reverted back into dictatorship, but Europe is also at the edge. Shouldn't we rather fight that nationalistic power grab that just makes us all poorer and less free? And instead propagate global alternatives that are not subjected by some power-hungry state-/capital-sponsored overlord?</p>
]]></description><pubDate>Fri, 23 Jan 2026 16:02:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=46734075</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=46734075</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46734075</guid></item><item><title><![CDATA[New comment by loehnsberg in "American importers and consumers bear the cost of 2025 tariffs: analysis"]]></title><description><![CDATA[
<p>This is a perfect example of how to lie with statistics. All of these countries are either tax havens or oil-rich economies, apart from half of them having the population of a small city. The economic policy implemented by any of these countries cannot be implemented by a large economy with little or no natural resources, or would you recommend to Germany or Japan to just "HAVE" oil or open their banks as offshore foreign accounts?</p>
]]></description><pubDate>Mon, 19 Jan 2026 17:22:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=46681733</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=46681733</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46681733</guid></item><item><title><![CDATA[New comment by loehnsberg in "I replaced Windows with Linux and everything's going great"]]></title><description><![CDATA[
<p>Why Bedrock? Get the kids a Steam Deck, Prism Launcher, open a local server and boom :) It‘s not iOS convenience but they‘ll sure love to tinker with all the mods you can install.</p>
]]></description><pubDate>Sat, 10 Jan 2026 17:23:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=46567715</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=46567715</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46567715</guid></item><item><title><![CDATA[New comment by loehnsberg in "AI Slop Report: The Global Rise of Low-Quality AI Videos"]]></title><description><![CDATA[
<p>Haha… no bot here. Been using Kagi for years now. Not sure what you‘re searching for. My own tests, admittedly early, found no instances, where Google gave better results.</p>
]]></description><pubDate>Sun, 28 Dec 2025 10:30:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=46410055</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=46410055</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46410055</guid></item><item><title><![CDATA[So You Want Continuous Time Zones (2017)]]></title><description><![CDATA[
<p>Article URL: <a href="https://qntm.org/continuous">https://qntm.org/continuous</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=45703110">https://news.ycombinator.com/item?id=45703110</a></p>
<p>Points: 1</p>
<p># Comments: 1</p>
]]></description><pubDate>Sat, 25 Oct 2025 11:47:54 +0000</pubDate><link>https://qntm.org/continuous</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=45703110</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45703110</guid></item><item><title><![CDATA[New comment by loehnsberg in "Fastmail desktop app"]]></title><description><![CDATA[
<p>Indeed what I do as well; gives me apps for Youtube, Netflix, etc. The only downside is that you have to login if you do not use the "app" for a while. Would Electron get around this?</p>
]]></description><pubDate>Mon, 13 Oct 2025 07:10:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=45565606</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=45565606</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45565606</guid></item><item><title><![CDATA[New comment by loehnsberg in "Starship's eleventh flight test"]]></title><description><![CDATA[
<p>Suppose you have a liberal mindset and work there, you must bend the knee and practice anticipatory obedience, or why else would you tell the world that the rocket will be ”dropping into the Gulf of America?“</p>
]]></description><pubDate>Sat, 04 Oct 2025 07:18:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=45471310</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=45471310</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45471310</guid></item><item><title><![CDATA[New comment by loehnsberg in "The best YouTube downloaders, and how Google silenced the press"]]></title><description><![CDATA[
<p>It‘s still a stark abuse of power and borderline extortion by Google to use a private sentencing mechanism rather than dragging the purpetrator to public court over advertising and/or encouraging criminal activity, which may or may not have happened if Google Ads and Youtube were not part of the same monopolistic entity.</p>
]]></description><pubDate>Sat, 20 Sep 2025 11:13:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=45312295</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=45312295</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45312295</guid></item><item><title><![CDATA[New comment by loehnsberg in "DeepSeek writes less secure code for groups China disfavors?"]]></title><description><![CDATA[
<p>Did they use the online Deepseek Chat or the open source model. If you ask either about the Tianenmen Square you get very different answers, which may be true for response quality as well.</p>
]]></description><pubDate>Wed, 17 Sep 2025 19:46:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=45280550</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=45280550</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45280550</guid></item><item><title><![CDATA[New comment by loehnsberg in "LLM leaderboard – Comparing models from OpenAI, Google, DeepSeek and others"]]></title><description><![CDATA[
<p>Interesting to learn that o4-mini-high has the highest intelligence/$ score here at par with o3-pro which is twice as expensive and slow.</p>
]]></description><pubDate>Fri, 01 Aug 2025 08:43:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=44754363</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44754363</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44754363</guid></item><item><title><![CDATA[New comment by loehnsberg in "Signal – An Ethical Replacement for WhatsApp"]]></title><description><![CDATA[
<p>I made that same experience. You can get individual people to install it, sure, you may even get a group to do this if you start it, but good luck convincing the other parents of your kid‘s sports team group, which btw you must be part of.<p>Unfortunately, we living beings tend to go with what costs the least amount of energy - this being thinking and going through extra efforts to achieve a goal. Hence, we‘re stuck with WhatsApp by a law of nature.</p>
]]></description><pubDate>Sat, 21 Jun 2025 06:59:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=44335238</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44335238</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44335238</guid></item><item><title><![CDATA[New comment by loehnsberg in "Time Series Forecasting with Graph Transformers"]]></title><description><![CDATA[
<p>L1-regularized autoregressive features, holiday dummies, Fourier terms (if suitable in combination) yield lower test errors, are faster in training, and easier to cross-validate than Prophet.</p>
]]></description><pubDate>Tue, 17 Jun 2025 20:55:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=44303852</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44303852</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44303852</guid></item><item><title><![CDATA[New comment by loehnsberg in "Last fifty years of integer linear programming: Recent practical advances (2024)"]]></title><description><![CDATA[
<p>Check out the Mittelmann benchmarks: <a href="https://plato.asu.edu/bench.html" rel="nofollow">https://plato.asu.edu/bench.html</a></p>
]]></description><pubDate>Sun, 15 Jun 2025 08:42:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=44281284</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44281284</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44281284</guid></item><item><title><![CDATA[New comment by loehnsberg in "Last fifty years of integer linear programming: Recent practical advances (2024)"]]></title><description><![CDATA[
<p>I would argue that not spending money on it and showing to upper mgmt that the folks they hired can actually get the job done often contributes to an external contractor not getting hired.<p>You might want to question, why didn‘t you ask the guys for money before starting to work for them. True, but I guess they were of the kind, show me results and then maybe we move on. On the other side, a 30-40k pilot project in this area is not difficult to negotiate if you‘re patient.<p>It takes so much more to running a business than lower cost with clever math that this step often comes at a later stage when larger companies look for ways to stay competitive, which is when they start to take a look at their accounts and figure that certain cost really stack up. Then you come in. The only real power you would have gotten over that company would have been for those guys getting fired and replaced by a vendor - ideally that‘s you!</p>
]]></description><pubDate>Sun, 15 Jun 2025 08:39:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=44281272</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44281272</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44281272</guid></item><item><title><![CDATA[New comment by loehnsberg in "Maximizing Battery Storage Profits via High-Frequency Intraday Trading"]]></title><description><![CDATA[
<p>Assuming you add all the annoying details that algo trade execution brings, the algorithm still provides the answer on which position to take within a few microseconds, which is what you want if you trade in a limit order book.</p>
]]></description><pubDate>Thu, 12 Jun 2025 19:24:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=44262076</link><dc:creator>loehnsberg</dc:creator><comments>https://news.ycombinator.com/item?id=44262076</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44262076</guid></item></channel></rss>