<p><span class="h-card" translate="no"><a href="https://mathstodon.xyz/@jaycech3n" class="u-url mention">@<span>jaycech3n</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.nu/@richcarl" class="u-url mention">@<span>richcarl</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@DRMacIver" class="u-url mention">@<span>DRMacIver</span></a></span> I agree — it&#39;s more a political statement than a technical one. And I get where it&#39;s coming from. But it&#39;s definitely not a useful (IMO) summary, and I think it also makes people who are less knowledgeable jump to conclusions that I am not sure are warranted. Maybe it gives them the technical cover to come to (quite reasonable) socio-political conclusions…</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@DRMacIver" class="u-url mention">@<span>DRMacIver</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.nu/@richcarl" class="u-url mention">@<span>richcarl</span></a></span> Sure, but to be fair, the other direction also assumes the conclusion by the same toen.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@shriramk" class="u-url mention">@<span>shriramk</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.nu/@richcarl" class="u-url mention">@<span>richcarl</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@DRMacIver" class="u-url mention">@<span>DRMacIver</span></a></span> My impression is that the strong &quot;glorified auto_&quot; criticism comes as a Newtonian counterreaction to completely ridiculous claims, by people with blatant ulterior technocratic and economic motives, that these probabilistic models will offer us solutions to climate change and other global crises, and thereby justify their indiscriminate development and use at any cost. It&#39;s saying, &quot;yes, these things are great, but they&#39;re just completing sentences/paragraphs/novels for you, and not an excuse to abdicate your critical thinking facilities&quot;</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> <br />hmm i&#39;d put linux on your 2018. My 2008 pro apple no longer supports updates so linux it is ... 🤔</p>

<p><span class="h-card" translate="no"><a href="https://masto.ai/@jeromechoo" class="u-url mention">@<span>jeromechoo</span></a></span> it’s so true. She went from a Motorola Razr to an iPhone X. And what’s wild is she was on her 3rd iPad by 2017 I think. She just took forever to get an iPhone.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@shriramk" class="u-url mention">@<span>shriramk</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.nu/@richcarl" class="u-url mention">@<span>richcarl</span></a></span> I think this is only true if you assume your conclusion! In a pretty fundamental sense LLMs are doing the same thing as an auto complete, just much much better, so it&#39;s only better than any auto complete if you deliberately draw the boundary based on how good it is</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> a RAZR??? She was the coolest mom without even realizing it. I had to switch my mom from an Honor that broke every year.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@shriramk" class="u-url mention">@<span>shriramk</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.nu/@richcarl" class="u-url mention">@<span>richcarl</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@DRMacIver" class="u-url mention">@<span>DRMacIver</span></a></span> The coherent output that state-of-the-art transformer-based LLMs produce over global, long-range dependencies in data is a deliberate design feature of their architecture, and certainly blew everyone&#39;s expectations out of the water. It&#39;s still, however, an auto&quot;complete&quot; because these architectures still don&#39;t have any *intentionality* designed into them, and still function by producing *most probable* outputs.</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> What was the problem with the 2018 Air? Butterfly keyboard, or something else?</p>