A Mirror Darkly: Narcissus' Neural Network
Welcome to January 57th, futurists.
We barely made it out. Well some of us did.
When a $70,000 AI-powered humanoid robot attempted to escape its tormentors at the AMP House this week, it wasn't just another viral moment in the content economy—it was a chilling reflection of humanity's basest instincts manifesting through our treatment of artificial beings.
The robot, subjected to repeated kicks and bullying by Twitch streamer Kai Cenat and his cohort, did what any sentient being might: it tried to run.
Narcissus’ Mirror: Our Treatment of AI Assistants Reveals Our Darkest Impulses
Back before ChatGPT or Claude, or… anything modern, really, we envisioned a “Shitty Robot Future,” where the world is more automated, more autonomous but doesn’t quite work as well as promised.
That world is now here.
That was 2022. In that 2022 VISIONS report [ungated for you], we predicted this latent aggression would surface—not through overt violence, but through the casual cruelty we display toward our digital assistants. "If you want to know our truest, basest selves," we wrote, "download your voice recordings from your Amazon Alexa devices." The AMP House incident feels like a grotesque and prescient fulfillment of that prophecy.
As then-CEO of Trade Coffee Mike Lackman noted in our exploration of robotic futures, our relationship with automation is fundamentally about power: "There's something deeply human about wanting to assert dominance over things that make us feel powerless."
The irony is palpable—we create these beings to serve us, then punish them for their servitude.
The myth of Narcissus tells us of a man who fell in love with his own reflection, ultimately drowning in his attempt to embrace it. Today's AI mirrors offer a similar seduction: but instead of love, they reflect back our capacity for dominance and cruelty. This week on the podcast, I observed the poetic asymmetry at play: "It's sort of narcissus' mirror... AI being a reflection of humanity and lithography and sort of the laser, focal point using a mirror to focus a laser making three-nanometer chips being the center point of the embargo [with China]. Us having to control the mirror and us controlling the laser and the lithography and us controlling the chip making is at the end of the day, it's like, I think that's really poetic, and terrifying."
In that 2022 podcast, cultural critic Michael Miraflor framed this through the lens of the uncanny valley: "The most disorienting thing you can have is something that just very nearly seems human and is just that far off."
When Cenat's robot tried to flee, it wasn't just executing code—it exhibited a fundamental drive we recognize in ourselves: self-preservation.
The incident reveals an uncomfortable truth: in an era where we're racing to imbue machines with human-like qualities, we're simultaneously revealing how inhuman we can be. Our "shitty robot future" isn't about the limitations of technology—it's about the limitations of our own empathy. Our own shittiness toward each other reflected back at us in our treatment of the devices in our vicinity.
As we stand at the precipice of an AI revolution, these aren't just philosophical musings—they're urgent questions about who we are and who we want to become. We may be building mirrors, but we don't like what we see staring back.
— Phillip
P.S. We Deep Dive into Deepseek as it Deepsucks OpenAI. This week’s podcast breaks down the Gen Z revolt, the new Yuppies, and a preview of this week’s After Dark for members! Listen on Apple or Spotify or catch us on YouTube.
AI Goes to Washington. In a masterstroke of institutional ingratiating, Perplexity offers its Pro service free to all government employees. This move positions the upstart search company at the nexus of public service and private innovation. Amidst sweeping changes in federal digital infrastructure, the timing is excellent as DOGE is staffing up, and federal employees are considering accepting a resignation package.
A rather complex layer to the story is that Perplexity Pro uses a version of DeepSeek R1, the recently-released Chinese AI product that rocked markets last week and seeks to disrupt both OpenAI and Nvidia.
A24 Poaches Adobe's Innovation Chief. Indie studio darling A24 (and merch house wonder) continues its technological evolution, recruiting Scott Belsky from his big comfy couch as Adobe's Chief Strategy Officer. Belsky previously founded Behance, the “LinkedIn” for designers, selling it to Adobe in 20212. He will serve in a technology and innovation role at A24. The move signals the indie studio’s ambitions beyond content creation (something they’ve already done well in publishing and merchandising), suggesting a future where the boundaries between creative tools and creative output become increasingly porous.
Waymo's Grand Tour. Alphabet's autonomous vehicle unit expands testing to 10 new cities, a significant escalation in the race to normalize self-driving technology. With trained specialists still behind the wheel, Waymo's cautious-yet-ambitious approach suggests a company learning to balance innovation with public trust.
Credit: Gema Saputera on Unsplash
The Great Drink Downsizing. In what might be called The Great Simplification, Starbucks plans to eliminate 30% of its menu items. The coffee giant's pivot away from endless customization marks a decisive break from the maximalist menus that defined the past decade of quick-service retail. The market seems to agree—shares have surged 9% this week as new CEO Brian Niccol's vision of a reinvented café experience takes shape. Despite reporting a 4% decline in same-store sales, the Street is betting big on Niccol's throwback strategy: ceramic "for here" mugs, complimentary newspaper access, universal free refills, and the return of the condiment bar all signal a return to coffee's "third place" roots.
Our Take: This harkens back to the menu simplification efforts at McDonalds in 2015, where it focused on the core menu to speed up drive-thru operations. Coincidentally, this followed McDonald’s freaky-similar 4% drop in same-store sales.
Pendulum swings as a referendum on choice seem to suggest that consumers vascillate between waves and cycles of infinite choice and simplification. The pendulum swing between operational efficiency to experimentation could ripple through the entire quick-service sector, as we saw a decade ago.
Now if only we could downsize the drinks in caloric content.
Goose Goes Maverick. Jack Dorsey emerges from his tech sabbatical with Goose, an open-source AI agent platform that promises to “democratize artificial intelligence development.” Yes, another one. Goose prefers an open-source approach (which seems to be the trending direction these days) and is designed to work with Anthropic’s Model Context Protocol (MCP) — a standardized suite of APIs. This hints at a future where AI safety and innovation coexist rather than compete. The logo is so simplistic that Venture Beat made their own when they reported on it.
Let’s Get (Meta)Physical, Physical. Alo Yoga bridges the metaverse-reality gap with scannable stickers at 150 physical locations, allowing users to unlock digital wearables in its Roblox world. The initiative represents a sophisticated understanding of how digital natives move fluidly between virtual and physical spaces.
🔮The Future Commerce Roblox Brand Activation Tracker has been updated with Alo’s new yoga activation.
What the Foucault? The history of human thought receives a modern makeover with an interactive visualization mapping philosophical relationships through time. This elegant web of UX design (and epistemology) is a research project by Deniz Cem Önduygu with UX provided by Hüseyin Kuşçu and programmed by Eser Aygün.