실리콘 밸리에서 혼돈의 위험이 만연함에 따라 AI 에이전트는 미트백을 고용합니다.

As technology journalists, we often come across a lot of frankly absurd startups. From subscription-based juicers to offers of worthless crypto in exchange for scans of your eyeballs via a metal orb, it's hard to imagine just how inane and insane some so-called founders think. A new challenger approaches. Enter RentAHuman, a platform that AI agents can use to find humans to perform actions they themselves are not capable of completing. Think about that for just a second. It's the Silicon Valley equivalent of hiring day laborer from the Home Depot parking lot. And if the startup's website is to be believed, they've got more than 600,000 "humans" signed on looking to be hired. At least they don't have the fear of being deported. This Fiverr meets "I, Robot" marks the next devolution in the race to the bottom. The launch of ChatGPT has led many in the tech space to sit back and ponder: How can I make things worse? Of course, this isn't limited to the startup space, with Microslop CEO Satya Nadella asking the world to essentially stop being mean to AI as it throws billions on the buildout bonfire. So why is RentAHuman so triggering to this hack journalist? It's a synonymous moniker of the stupidity we're living in. Where the inherent desire to slap the label of "the next big, hot thing" outweighs the very idea of what innovation is about. The concept of an AI agent – whereby a system is capable of performing a predetermined task or action(s) with little to no human intervention – is a natural next step in the evolution of the plain old chatbot. Up until now, of course, the underlying model components simply haven't been up to snuff. But with each passing year comes iteration and architectural innovation that's genuinely impressive. For all the saber-wrattling around DeepSeek, for example, it shows you can do a lot in AI with not much hardware. The folks over at Anthropic came out of OpenAI and (with the help of a few billion dollars from Amazon and Google) created what some believe to be a far superior product. But the race to agentify everything raises the all too familiar face of FOMO (fear of missing out) ... on showing you're hip and with it. For whatever you think of Gartner, their report on "agent-washing" was a breath of fresh air for an industry drunk on its own hubris. Where one might cry agent, the reality is a simple RPA application repackaged. Like Alan Partridge's Rover Metro, "they've rebadged it, you fool." RentAHuman, then, is everything AI shouldn't stand for. What on Earth is the point in putting in place a system in which at least part of its tasks vital to its function cannot be performed by the system? And pray, if you were to go onto this platform and look at the dross being advertised as skills, it's laughable. There’s folks offering to hold signs in public places for $25 an hour, solving Rubik’s Cubes in under a minute, and potentially worrying, to “play with you.” Heck, there are even users offering multiple different services. Like this guy, who does everything from creating custom music tracks to providing recordings of human point-of-view (POV) footage for data annotation. And why would an AI agent need a 30-minute “Business and life clarity session?” But the cherry on the icing atop the "I told you so" cake? Allegedly, real human users are offering to create AI-generated images for those agents to use. AI agents, hiring a human worker to go to another AI and ask it to create a picture. I've come across a lot in my time, from a "friendship" chatbot created using capital from Russian mobsters to supposedly autonomous humanoid robots that poorly paid people in India actually pilot. Vtuber singers, AI muscians, and morons still trying to make the flying car happen, the tech industry is filled with snake oil salespeople. Just run that idea back one last time: You implement an AI agent to perform a task. For some reason, you've decided to deploy an agent for a task that requires images despite it being incapable of creating or sourcing said images. So instead of putting in a plug-in to DALL-E or any number of image generating models, or providing it with access to a pre-existing licensed content, it opts to "hire" a human, who then simply goes to ChatGPT, types in a prompt – and that's assuming the AI doesn't gives the human a prompt, in what would eliminate the only actual potential for cognitive processing on the human worker's part – and make them hand it back in exchange for chump change. That alone spits in the face of innovation and instead perpetuates the classist nomenclature we've learned to accept in society. There's more dignity going down a mine for Pete's sake. So while I'm lambasting a startup, let me balance with a company that's looking at AI the right way: in that they want nothing to do with it: Games Workshop – the makers of "Warhammer." Now, long-time SDxCentral readers who have made it this far down are likely sighing right now. But I promise this isn’t another attempt by this editor to write about yet another nerdy pastime following a recent opus on Pokémon cards. But Games Workshop's stance on AI has been refreshing – value human creators, with AI instead a cheap imitation slop machine (sorry, Satya). During the company’s recent earnings, CEO Kevin Rountree said a few of its senior managers were experimenting with AI, but “none are excited about it yet.” “We have agreed an internal policy to guide us all, which is currently very cautious, e.g., we do not allow AI-generated content or AI to be used in our design processes or its unauthorized use outside of Games Workshop, including in any of our competitions," Rountree said. There's an irony in the makers of "Warhammer" disliking AI, when the entire backstory for "Warhammer" 40K's grimdark future hinges on such a technology running amok and causing humanity to be fractured and splintered, which in turn opens them up to the perils of chaos. Spooky. In fact, creating a "Silica Animus" or "Abominable Intelligence" is a crime under the laws of the Imperium of Man by order of the Emperor himself. The AI distrust from "Warhammer’s" own creators is, however, from a position of power, in that it wants to control its own intellectual property. The British company has licensing practices that would make the lawyers at Disney shudder. And only recently, and likely reluctantly, letting Amazon play with its toys for a TV series starring Henry Cavill, given that it offers its own streaming content. But their stance on AI is the antithesis of everything RentAHuman stands for: letting humans do real, actual, creative work, and not bit part tasks akin to selling your retinal scans to the folks at Worldcoin. It's a welcome change from the agents that are everywhere dominating market discourses right now. Comments