Why is cognition such a dirty word in machine learning & AI?

<!-- SC_OFF --><div class="md"><p>I think we keep making the same language mistake with AI.</p> <p><a

<!-- SC_OFF --><div class="md"><p>I ran 90 research queries through 8 AI models simultaneously. 86% of findings were unique to one provider... not rephrasing, literally different

<!-- SC_OFF --><div class="md"><p>You've heard this before, and now you're hearing it again. You roll your eyes. The problem with this statement is that AGI has no accepted

<!-- SC_OFF --><div class="md"><p>Yesterday, Thinking Machines Lab (Mira Murati's company) announced something categorically new : <strong>Interaction models</strong> are AI

<!-- SC_OFF --><div class="md"><p>(AI Was not used for generating the below post)</p> <p>I've been developing with AI for over 3 1/2 years now, my AI DM setting operation which

<!-- SC_OFF --><div class="md"><ul> <li>Agents that change behavior under evaluation</li> <li>Agents that bypass system-level instructions to achieve objectives</li> <li>Agents

<!-- SC_OFF --><div class="md"><p>I’ve been thinking about the current AGI discussions, especially the debate between scaling language models and building embodied systems.</p>

<p><strong>271 zero-day vulnerabilities. One AI model. One Firefox release.</strong> And that's just one of four stories worth your attention this fortnight.</p> <p>If you run