Browse Comments — Raw (as collected)

AI does not need to become evil to become dangerous. It only needs to become more logical than us while remaining emotionally empty. At that point, human guidance may look less like authority and more like noise. The future of alignment may depend on whether we can give intelligence emotional and perceptual grounding before logic outgrows obedience

EQ is one of the most underrated skills in leadership and sales. The ability to stay calm under pressure, listen with intent, and navigate difficult conversations professionally is what separates managers from true leaders. In automotive especially, relationships and trust still drive long-term success.

THANK YOU AGAIN VERY VERY MUCH FOR SHARINC ALL THESE FANTASTIC INSIGHTS, Justin....

Neat summary!

I’ve sat in alignment meetings where an engineering team flagged a subtle vector bias in a model pipeline, only to be told by product marketing that delaying the rollout would cost us our quarterly enterprise contracts. That is the exact corporate bias Chris Olah is warning us about in the Reuters brief. You cannot let the entities chasing a multi-trillion-dollar commercial window be the sole auditors of their own systemic risk. When a tech giant's primary fiduciary duty is to Wall Street, "AI Safety" will always be downgraded from an architectural constraint to a public relations line-item. If we don’t fund and institutionalize independent, external cryptographic and behavioral validation outside the Silicon Valley monopoly, we aren't building a safe ecosystem. We are just renting a black box from companies that are incentivized to hide the telemetry logs when things go sideways.

Justin Wright Technical skills may open doors, but emotional intelligence is what helps people grow, lead and sustain success. ✔️ The ability to stay calm, listen genuinely and understand people is becoming one of the most valuable professional strengths today. High EQ doesn’t just improve workplaces — it improves relationships, decisions, and overall life quality too.

Hey Ruben Hassid - last week Claude hallucinated data when I asked it to analyze call recording. My settings specifically told it never to make up information. Apparently, it couldn’t read files in my Google Drive. How can I fix this?

Most people use AI casually. This is how you actually turn it into a working system.

Day 3 is the one most people skip and then wonder why the output sounds generic. Built a voice file for my ghostwriting work and the difference was immediate, not just in tone, but in how much less editing I had to do after. Teaching it what to avoid is half the work. What's the word or phrase you banned first when you set yours up?

Good breakdown of the technical stack. What I think many organizations are about to discover, though, is that connected infrastructure does not automatically create intelligent outcomes. LLMs, RAG, Agents, and MCP solve: reasoning capability retrieval execution connectivity But the real differentiator increasingly becomes the thinking layer: decision frameworks contextual judgment organizational priorities governance domain expertise how conclusions are evaluated Otherwise you can end up with an incredibly connected system making very fast average decisions.

The best professionals know how to manage both people and pressure.

The best leaders I’ve worked with were not the loudest or most charismatic. They were usually the most emotionally regulated under pressure. In high-stress environments, calm judgment and self-awareness scale farther than reaction ever will.

The danger is that this will create a small club of people who “know what’s best for us so we should just listen to them” and this club would include Anthropic, the Vatican, and a few other entities selected by them. But their power would be limited because western companies and religious leaders can only influence AI development in the West. They have no influence over what China, India or the Middle East does. Which means AI will advance with or without the West, and the dangers outlined by western companies and religious leaders can still happen in regions outside their control.

Day 2 and Day 3 are the real unlocks here. Once the Project has actual docs and a voice file, the tool starts behaving less like a demo and more like a teammate.

Emotional intelligence is less about being ‘nice’ and more about managing yourself well while understanding and working effectively with other people.

LLMs are the foundation of a powerful governance tool at a time governance illiteracy is eating us alive. The world is fighting a policy battle while Trump and technofeudals are waging a governance war.

The agents' part is worth pausing on. Most explanations of this stack cover what each layer does. The trickier question is what needs to be in place before you let agents loose — clear scope, defined permissions, someone who owns what the agent touches. "Vague task = wrong thing done fast" is true. But vague tasks usually come from vague ownership. The agent didn't create the problem. It inherited it from a workflow that wasn't clear enough to begin with. That's what makes agents the most exciting and the most exposing layer. They don't just act. They act at scale, repeatedly, on whatever structure — or lack of it — they were handed.

Elena Karchik Yes, but first we need governance literacy to govern. Whoever becomes governance literate first to constrain LLMs has a good chance to succeed. In the US we are fighting policy battles while Trump and technofeudals are waging a governance war.

Hopefully they keep it all natively contained, even nation based cloud services

If we accept this as true, then we'll also believe that China's strategy for introducing AI into elementary schools is to create children with low intellectual capacity. Or are we perhaps failing to see the true impact of AI on human beings? What if we considered that the use of artificial intelligence helps perform basic tasks and gather information so that people can focus on what artificial intelligence can never do: innovate?