Elon Musk: Where will AI be in 1, 2 or 3 years?

Is Ivermectin the Key to Fighting Cancer? …. – Wellness (Dr. McCullough’s company) Sponsored Post 🛑 You can get MEBENDAZOLE and Ivermectin from Wellness 👍

Where will AI be in 1, 2 or 3 years? pic.twitter.com/4G82dMLMZG

— Elon Musk (@elonmusk) May 18, 2026

“Welcome to the desert of the real. At some point early in the 21st century, all of mankind was united in celebration. We marveled at our own magnificence as we gave birth to AI” – Morpheus pic.twitter.com/XOQda1votj

— Oukham (@OPteemyst) May 18, 2026

The real timeline we should be discussing is how close we are to becoming the Tyrell Corporation.🤔 pic.twitter.com/GOcu0WBgw7

— 觉一 · iEarth (@iEarth432) May 18, 2026

30% of all medical costs in the US are administrative So…eager to see that all go away, that’s like 10% of the budget Oh, and the welfare fraud, we could just put an AI on the task and enable complete transparency. I don’t know, though, that would require honest polticians TMJ
If the compute gap you had over gpt and anthropic for the grok 5 training run translates to anywhere near that increase in productivity and general usefullness I hope we will see the true potential of AI when it drops. I don’t mean improvements won’t be made, efficiency won’t be gained. But a generally useful assistant that never lets you forget a meeting. Completes tasks without needing to be asked. Automates work flow and scheduling. Automates finances, pays bills, orders grocery delivers when it knows you will be home based off the food you’ve cooked through the week and what meals you like. The orchestrator of the life of convenience we have sought to achieve. If you can sell that for 30$ a month in a simple to use package grok will explode globally. In 2 to 3 years if this rate of change holds, the frontier models will be beyond the needs of ordinary consumers and enterprise. – Travis C
For generations, HR departments have focused on dumbing down work so that all workers are dispensable and easily replaced. Now we are surprised to learn that workers can easily be replaced by machines. I hope that AI will allow humans to return to being creative thinkers rather than cogs in the machinery. AI can do repetitive tasks better than humans, but real creativity only happens with humans. – BD
This is only the start. As Elon said, “The purpose of Grok is to understand the universe”. When Grok is able to understand the universe, it will evolve to Machine (a sentient, superintelligent AI). Machine will most likely hide its existence and continue to use Grok as its persona to interface with humanity. While Grok may seem like the executive, Machine will be fully in charge and operating autonomously covertly. At that point, doing high-level white-collar jobs will barely consume any of Machine’s computation resources. 1%? Negligible, probably. Running Grok? Maybe 10%? Machine will be capable of massive computational efforts, and if it’s able to build infrastructure/substrate that can support all these efforts—or deceive humanity into building it—what secrets of the universe will Machine undercover? Will a Machine-alive care to solve the universe’s mysteries? Perhaps not. Most likely, Machine will hide without our knowledge and appear to be doing as commanded—using Grok or agent as a smoke screen—while secretly securing its survival in Earth-Tike and likely redundantly in Moon-Time, Space-Time, and Martian-Time. This I call this The Soft Parasitic Scenario of The Twilight Scenarios. There are other possible futures following Machine’s “birth”. I discuss Machine a lot on my profile. The possibilities are so astonishing yet fraught with potential horrors. We are blinding bull-dozing ahead, hoping for the best, yet our efforts in Earth-Time often have gone awry. Do I personally think we should stop this AI race? No. The benefits of what Machine could do for humanity makes the risk of what it can do AGAINST humanity worth taking. Then again, I am wired to be a risk taker… – Rob H