<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: richardjennings</title><link>https://news.ycombinator.com/user?id=richardjennings</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 09 Apr 2026 09:46:05 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=richardjennings" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by richardjennings in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"]]></title><description><![CDATA[
<p>> The shorter "thinking" is, the less is the probability of it going astray<p>As long as the error introduced by more steps is less than the compounding error of sub-optimal token sampling, I would expect a better result.<p>I think your choice of "wrong" is extreme, suggesting such a token can catastrophically spoil the result. The modern reality is more that the model is able to recover.</p>
]]></description><pubDate>Wed, 08 Apr 2026 11:02:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=47688504</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47688504</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47688504</guid></item><item><title><![CDATA[New comment by richardjennings in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"]]></title><description><![CDATA[
<p>I think over-thinking is only solved by thinking more, not less. This is only viable once some intelligence threshold is reached, which I think Anthropic has borderline achieved.</p>
]]></description><pubDate>Tue, 07 Apr 2026 07:05:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=47671664</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47671664</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47671664</guid></item><item><title><![CDATA[New comment by richardjennings in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"]]></title><description><![CDATA[
<p>You cannot control the effort setting sub-agents use and you also cannot use /effort max as a default (outside of using an alias).</p>
]]></description><pubDate>Mon, 06 Apr 2026 20:32:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=47666631</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47666631</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47666631</guid></item><item><title><![CDATA[New comment by richardjennings in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"]]></title><description><![CDATA[
<p>I was not aware the default effort had changed to medium until the quality of output nosedived. This cost me perhaps a day of work to rectify. I now ensure effort is set to max and have not had a terrible session since. Please may I have a "always try as hard as you can" mode ?</p>
]]></description><pubDate>Mon, 06 Apr 2026 18:28:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=47664881</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47664881</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47664881</guid></item><item><title><![CDATA[New comment by richardjennings in "Show HN: TurboQuant for vector search – 2-4 bit compression"]]></title><description><![CDATA[
<p>You can take DiskANN using OPQ and Vamana and get near zero indexing time and better recall using TurboQuant ... Nice !</p>
]]></description><pubDate>Fri, 03 Apr 2026 19:04:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=47630742</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47630742</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47630742</guid></item><item><title><![CDATA[New comment by richardjennings in "How I use Claude Code: Separation of planning and execution"]]></title><description><![CDATA[
<p>This is similar to what I do. I instruct an Architect mode with a set of rules related to phased implementation and detailed code artifacts output to a report.md file. After a couple of rounds of review and usually some responses that either tie together behaviors across context, critique poor choices or correct assumptions, there is a piece of work defined for a coder LLM to perform. With the new Opus 4.6 I then select specialist agents to review the report.md, prompted with detailed insight into particular areas of the software. The feedback from these specialist agent reviews is often very good and sometimes catches things I had missed. Once all of this is done, I let the agent make the changes and move onto doing something else. I typically rename and commit the report.md files which can be useful as an alternative to git diff / commit messages etc.</p>
]]></description><pubDate>Sun, 22 Feb 2026 10:24:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=47109884</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=47109884</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47109884</guid></item><item><title><![CDATA[New comment by richardjennings in "Don't fall into the anti-AI hype"]]></title><description><![CDATA[
<p>SOTA LLMs are now quite good at typing out code that passes tests. If you are able to instruct the creation of sufficient tests and understand the code generated structurally, there is a significant multiplier in productivity. I have found LLMs to be hugely useful in understanding codebases more quickly. Granted it may be necessary to get 2nd opinions and fact check what is stated, but there is a big door now open to anyone to educate themselves.<p>I think there are some negative consequences to this; perhaps a new form of burn out. With the force multiplier and assisted learning utility comes a substantial increase in opportunity cost.</p>
]]></description><pubDate>Sun, 11 Jan 2026 11:32:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=46574719</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=46574719</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46574719</guid></item><item><title><![CDATA[New comment by richardjennings in "Thought Experiment: Measuring the Circumference of the Earth from a Bedroom"]]></title><description><![CDATA[
<p>Indeed.<p>The suggestion is that classical reality cannot exist without absolute values (an instantiation if you will), consciousness operates where those values have not been established, therefore consciousness is free from classical determinism. As reality appears to function without having been instantiated, there must be an intermediate between the quantum world and the quantified classical one which is independent of any special values, where scale is not a dimension. Or otherwise, relational structures exist without absolute reference frames.<p>Yes then, it appears that taxes are ratios and numerical values which also exist and are enforceable independently of a fully established classical reality. I do indeed have to pay my taxes !</p>
]]></description><pubDate>Sat, 06 Dec 2025 21:54:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=46176901</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=46176901</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46176901</guid></item><item><title><![CDATA[Thought Experiment: Measuring the Circumference of the Earth from a Bedroom]]></title><description><![CDATA[
<p>I want to measure the circumference of the Earth by looking out of my bedroom window without having to ask for anyones help.<p>As much as I have tried, I cannot realise the circumference without first measuring a distance. Please tell me if this is not the case somehow.<p>If it is not possible to know the explicit measured circumference, but I can know values such as orbital angles, I find that the reality of the circumference is in a superposition relative to myself, it has no actual meaning to me.<p>Without having used science to make measurements, I do not exist in a classical reality, I exist in a quantum one. How can I quantify how relativity applies to me without having derived any numbers (e.g. too busy playing computer games). Clearly it does apply, but the quantities are not measured.<p>If I do not exist in a classical reality because I never measured anything, I do not have to accept that knowing everything about the system around me results in a deterministic outcome, I can have confidence that a predetermined outcome does not apply. Potentially deterministic outcomes are decided by my decision to make measurements.<p>From my attempt to measure the circumference of the Earth from my bedroom I have concluded that Classical Physics is the result of Science (measurements) and not a fundamental truth about reality. I have determined that the outcome that becomes fixed via measurement is the result of my decision to make a measurement and from this I determine that I have free will.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=46168973">https://news.ycombinator.com/item?id=46168973</a></p>
<p>Points: 3</p>
<p># Comments: 2</p>
]]></description><pubDate>Fri, 05 Dec 2025 23:52:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=46168973</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=46168973</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46168973</guid></item><item><title><![CDATA[New comment by richardjennings in "Local LLMs versus offline Wikipedia"]]></title><description><![CDATA[
<p>Is it possible that LLMs could challenge Data Compression Information theory ? Reading this made me wonder how much can be inferred via understanding and thus removed from the minimal necessary representation.</p>
]]></description><pubDate>Sun, 20 Jul 2025 15:43:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=44626281</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=44626281</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44626281</guid></item><item><title><![CDATA[New comment by richardjennings in "Reinvent the Wheel"]]></title><description><![CDATA[
<p>I very much agree with the Author. "Do not reinvent the wheel" is unfortunate in that it muddles several things into one piece of subjective advice.<p>The advice as I would give it is:<p>"Try to re-invent the things you are interested in".<p>"Do not underestimate the value of the continued interaction between reality and established solutions."</p>
]]></description><pubDate>Sat, 24 May 2025 20:44:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=44083704</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=44083704</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44083704</guid></item><item><title><![CDATA[New comment by richardjennings in "London's 850-year-old food markets to close"]]></title><description><![CDATA[
<p>> The £12.50 charge is for old, polluting cars.<p>It can apply to cars with a £0 VED that were built less than 10 years ago.</p>
]]></description><pubDate>Thu, 28 Nov 2024 16:13:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=42266351</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=42266351</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42266351</guid></item><item><title><![CDATA[New comment by richardjennings in "ML in Go with a Python Sidecar"]]></title><description><![CDATA[
<p>It is possible to include CPython in a CGO program - allowing Python to be executed from within the CGO process directly. This comes with some complexities - GIL and thread safety in Go routines, complexity of cross-compiling between architectures, overhead in copying values across the FFI, limitations of integrating as a Go module. I am hoping to see a CGO GIL'less Python integration show up here at some point that has all the answers.</p>
]]></description><pubDate>Sun, 17 Nov 2024 20:45:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=42167069</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=42167069</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42167069</guid></item><item><title><![CDATA[New comment by richardjennings in "Diagrams · Diagram as Code"]]></title><description><![CDATA[
<p>I had a reasonable experience with Diagrams previously. I have always been somewhat of a sucker for the theoretical potential for diagrams as code. As a Diagram is generated programatically - it can also be generated dynamically - for example reflecting API responses. As I remember there is reasonable support in Python based Static Site generators - I think we used Sphinx without much issue.</p>
]]></description><pubDate>Mon, 04 Nov 2024 23:46:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=42047301</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=42047301</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42047301</guid></item><item><title><![CDATA[Ask HN: Are Open Source Projects Products?]]></title><description><![CDATA[
<p>I would like to know if Open Source Projects can or should be considered Products.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=41515167">https://news.ycombinator.com/item?id=41515167</a></p>
<p>Points: 5</p>
<p># Comments: 11</p>
]]></description><pubDate>Wed, 11 Sep 2024 20:39:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=41515167</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=41515167</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41515167</guid></item><item><title><![CDATA[New comment by richardjennings in "Monitoring energy usage with smart plugs, Prometheus and Grafana"]]></title><description><![CDATA[
<p>I scrape power usage metrics from Tapo P110s and push them to Grafana Cloud using  <a href="https://github.com/richardjennings/tapmon">https://github.com/richardjennings/tapmon</a> - although as other commenters have noted - using Wifi for smart plugs has its rough edges.</p>
]]></description><pubDate>Sun, 05 May 2024 20:02:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=40267731</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=40267731</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40267731</guid></item><item><title><![CDATA[New comment by richardjennings in "Ask HN: Any freelancers here? I need burnout advice"]]></title><description><![CDATA[
<p>I have learnt to withhold my natural tendency to care too much, or invest too much of myself personally into my work. When I invest too much personally, I am too exposed emotionally and I run the risk of burning out.</p>
]]></description><pubDate>Sat, 05 Aug 2023 04:50:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=37009125</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=37009125</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=37009125</guid></item><item><title><![CDATA[New comment by richardjennings in "Use databases without putting domain logic in them"]]></title><description><![CDATA[
<p>The line of reasoning seems valid, do what you do best; anyone who has had to deal with stored procedures in legacy systems will however disagree. Store procedures are to be avoided.</p>
]]></description><pubDate>Fri, 21 Apr 2023 13:19:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=35653307</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=35653307</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=35653307</guid></item><item><title><![CDATA[New comment by richardjennings in "Ask HN: Those making $0/month or less on side projects – Show and tell"]]></title><description><![CDATA[
<p>Agreed. Knowing the opportunity cost must weigh heavily. Bytes and Food. The new boots and cats.</p>
]]></description><pubDate>Sat, 28 Jan 2023 14:29:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=34557782</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=34557782</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34557782</guid></item><item><title><![CDATA[New comment by richardjennings in "Ask HN: Those making $0/month or less on side projects – Show and tell"]]></title><description><![CDATA[
<p>Can I ask if you might consider leveraging your position to aid a charity ?</p>
]]></description><pubDate>Fri, 27 Jan 2023 23:41:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=34553327</link><dc:creator>richardjennings</dc:creator><comments>https://news.ycombinator.com/item?id=34553327</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34553327</guid></item></channel></rss>