<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Quality on Programmer.ie: Modern AI programming</title>
		<link>http://programmer.ie/tags/quality/</link>
		<description>Recent content in Quality on Programmer.ie: Modern AI programming</description>
		<generator>Hugo</generator>
		<language>en-US</language>
		
		
		
		
			<lastBuildDate>Sat, 21 Nov 2026 10:00:00 +0000</lastBuildDate>
		
			<atom:link href="http://programmer.ie/tags/quality/index.xml" rel="self" type="application/rss+xml" />
			<item>
				<title>A New Approach to Quality</title>
				<link>http://programmer.ie/books/machine-coding/16-chapter/</link>
				<pubDate>Tue, 10 Nov 2026 00:00:00 +0000</pubDate>
				<guid>http://programmer.ie/books/machine-coding/16-chapter/</guid>
				<description>&lt;p&gt;Chapters 13 to 15 gave you the ability to create: build by speaking, think with the machine, judge the product rather than the code. This chapter asks the question those abilities make unavoidable. If you can build almost anything, what is worth building — and what makes the result any good?&lt;/p&gt;&#xA;&lt;p&gt;For most of the history of software, that question was partly answered by difficulty. Building was hard enough that the act of building something working was itself a distinction. That is changing. When competent software becomes something many people can produce, producing it stops being enough to set anyone apart. Something else has to.&lt;/p&gt;</description>
			</item>
			<item>
				<title>Benchmarking Against Established Engines</title>
				<link>http://programmer.ie/books/inference/44-chapter/</link>
				<pubDate>Tue, 20 Oct 2026 00:00:00 +0100</pubDate>
				<guid>http://programmer.ie/books/inference/44-chapter/</guid>
				<description>&lt;p&gt;The heavy workload supplies the chapter&amp;rsquo;s first warning against declaring a performance winner from one latency column. In the recorded five-trial comparison, the capstone with batched decode reached a median first-token time of 0.071 seconds; the local library loop using static batches of four reached 0.879 seconds. Their median output throughputs were nearly identical, at 56.05 and 55.28 tokens per second.&lt;/p&gt;&#xA;&lt;p&gt;Yet their median shares meeting &lt;strong&gt;both&lt;/strong&gt; service thresholds were 0.63 and 0.96, respectively. Getting the first token out sooner did not ensure that the rest arrived quickly enough. A repeated run under an identically configured capstone name reached a median SLO share of 0.91, an early warning that the apparent gap needs independent confirmation.&lt;/p&gt;</description>
			</item>
			<item>
				<title>The New Math</title>
				<link>http://programmer.ie/post/math/</link>
				<pubDate>Sat, 21 Nov 2026 10:00:00 +0000</pubDate>
				<guid>http://programmer.ie/post/math/</guid>
				<description>&lt;p&gt;&lt;em&gt;Quality Perception: The Human Skill at the Centre of AI-Assisted Research&lt;/em&gt;&lt;/p&gt;&#xA;&lt;p&gt;If you work with AI systems for long enough, you start to notice a rhythm.&lt;/p&gt;&#xA;&lt;p&gt;Most interactions are ordinary. The model does roughly what you asked. Some interactions are quietly worse than they look: an answer that is fluent but slightly off, a fix that passes the test by weakening the test, an explanation that has become longer without becoming clearer. And every so often, perhaps once in a few hundred exchanges, something exceptional happens. A connection appears that you didn&amp;rsquo;t ask for. A messy problem suddenly becomes simple. A result arrives that is better than anything you expected.&lt;/p&gt;</description>
			</item>
	</channel>
</rss>
