June 18 2026 - Goats, Skepticism & Robots: AI Reality CheckGoats, Skepticism & Robots: AI Reality Check

**Taylor** (0:00)
Welcome back to AI Signal & Noise. It is Thursday, and oh my gosh, we have some of the wildest stories today.
I'm Taylor, and I'm so excited for this episode.

**Morgan** (0:11)
And I'm Morgan. Wild is definitely the word today, Taylor. From virtual goats doing math to public skepticism, we're cutting through the hype to see what's real.

**Taylor** (0:22)
Oh, totally. We're talking real world robots, life science benchmarks, and literally goats, like video game goats. I am not even joking. It is so cool. Let's just dive right in.

**Morgan** (0:35)
I can't wait to hear how goats fit into AI research. Lead the way, Taylor, because I am highly skeptical of this one.

**Taylor** (0:43)
Okay, so this Microsoft researcher built an actual working neural network inside Age of Empires 2, using goats, bridges, and ice ramps. It is absolutely genius.

**Morgan** (1:00)
Wait, what? Goats? Like the little farm animals in a 20-year-old strategy game? How on earth does that even work?

**Taylor** (1:09)
Yes, he used the map editor. The goats wander around, and their paths simulate logic gates. It actually computes things, like addition. It's so incredibly cool, dude.

**Morgan** (1:22)
Okay, that sounds like a fun weekend project, but what is the actual point? Is this just a researcher with too much free time, or is there a real lesson?

**Taylor** (1:32)
It's actually a huge critique of AI science. He analyzed 315 research papers, and found more than half assume language models have human traits before testing.

**Morgan** (1:47)
Ah, anthropomorphism. We see this constantly in the industry. If you replace the chat box with wandering goats, the math is identical, but the illusion of consciousness completely vanishes.

**Taylor** (2:01)
Exactly. The goats show it's just pathfinding and logic. It doesn't mean the goat network understands anything, even if it gets the right answer in the end.

**Morgan** (2:12)
It's a brilliant reality check. It proves we are often measuring emergent behavior that we've practically programmed in, rather than actual human-like intelligence.

**Taylor** (2:23)
Right? Like, if a goat hits a switch, we don't say the goat knows calculus. It's just a great way to look at the math without the sci-fi hype.

**Morgan** (2:33)
Definitely. We need more of this grounding in the community.
What else did the researchers say about how we evaluate these models?

**Taylor** (2:41)
Well, he basically said we are testing them like they are humans, which is a huge mistake. We need to test them like the complex systems they actually are.

**Morgan** (2:51)
I completely agree. If we keep treating them like magic black boxes with feelings, we will never actually understand how to make them safe or reliable.

**Taylor** (3:01)
Totally. It really makes you rethink all those flashy AI benchmarks we see every single week, you know?

**Morgan** (3:09)
Speaking of how people perceive this stuff, that leads us perfectly into some pretty telling statistics about public trust.

**Taylor** (3:16)
Oh, totally! A new Pew research study just dropped, and dude, the numbers are absolutely wild. People are really not buying the AI hype train right now.

**Morgan** (3:28)
I saw that. Only 16% of Americans think AI will have a positive impact on society. That is incredibly low, especially given the billions being invested.

**Taylor** (3:39)
Right? Only 16%!
Like, Wall Street is going completely bananas over AI, but regular people are just like, uh, no thanks, we are good.

**Morgan** (3:50)
Well, can you blame them? People are worried about their jobs, misinformation, and frankly, the general weirdness of how fast this is moving without public consent.

**Taylor** (4:01)
Totally, but I wonder if it's just fear of the unknown. Like, once people use it more in their daily lives, won't they start to like it?

**Morgan** (4:09)
Actually, the study suggests the opposite. The more people hear about AI, the more anxious they seem to get about its integration into daily life.

**Taylor** (4:19)
That is so wild to me.
I use it every day and it saves me so much time. But I guess if you're worried about losing your livelihood.

**Morgan** (4:28)
Exactly. It's easy for tech enthusiasts to forget that for many, AI represents disruption and uncertainty, not just a cool productivity shortcut or a fun toy.

**Taylor** (4:40)
True. And the tech companies aren't really helping when they keep pushing things people didn't ask for, like AI and search or writing in Google Docs.

**Morgan** (4:49)
Right. Like those Gemini pop-ups in Google Docs that everyone is trying to turn off. It feels forced rather than helpful, which breeds resentment.

**Taylor** (4:59)
Oh, totally.
But seriously, if 16% is the ceiling right now, tech giants have a massive public relations problem on their hands. They need to listen.