<p>I’m not a lawyer, but if my company was currently being sued by the DOJ for anticompetitive behavior, I probably wouldn’t lie under oath in a closely-related case…</p><p>But maybe that’s just me! <a href="https://www.wired.com/story/antitrust-judge-asks-doj-prosecute-apple/" target="_blank" rel="nofollow noopener" translate="no"><span class="invisible">https://www.</span><span class="ellipsis">wired.com/story/antitrust-judg</span><span class="invisible">e-asks-doj-prosecute-apple/</span></a></p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@dabeaz" class="u-url mention">@<span>dabeaz</span></a></span> I&#39;m with you on that. These tools do the part I love (puzzles, design, problem solving) while leaving me with the parts I never wanted (admin, trying to get someone else to solve a problem instead of just doing it myself).</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@dabeaz" class="u-url mention">@<span>dabeaz</span></a></span> my biggest problem with LLM (especially &quot;agentic&quot; tool) is not even their buggy outputs or ethical concerns. Instead, it is that they make coding from something I enjoy into miserable experiences</p>

<p><span class="h-card" translate="no"><a href="https://infosec.exchange/@jann" class="u-url mention">@<span>jann</span></a></span> part of a CPU but i don&#39;t think it bumps into TDP limits... don&#39;t *think*</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@film_girl" class="u-url mention">@<span>film_girl</span></a></span> It probably also doesn&#39;t help that for those of us in the 180-ish other countries, AVP basically doesn&#39;t exist.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>whitequark</span></a></span> is the GPU part of a CPU that enforces a combined TDP limit or are those things completely separate?</p>

<p><span class="h-card" translate="no"><a href="https://infosec.exchange/@jann" class="u-url mention">@<span>jann</span></a></span> almost certainly yes</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>whitequark</span></a></span> does one of the two implementations involve more thread switches, where you have lots of "thread A wakes thread B, then thread A goes to sleep"? AFAIK the kernel already can't handle those particularly well (<a href="https://youtu.be/KXuZi9aeGTw?t=611" target="_blank" rel="nofollow noopener" translate="no"><span class="invisible">https://</span><span class="">youtu.be/KXuZi9aeGTw?t=611</span><span class="invisible"></span></a>), and I imagine adding more noise to the scheduler could make things worse?</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@dabeaz" class="u-url mention">@<span>dabeaz</span></a></span> Like you, I&#39;ve eschewed LLMs. They strike me as the automated equivalent of the lazy tutor, who shows you solutions to your homework problems without trying to improve your understanding.</p><p>So you get help with *this* problem, but all you&#39;ve learned in terms of tackling the rest of the problems you&#39;ll encounter is dependency.</p>