Evidently we donât have to âreinvent the wheelâ to find new ways to do things (or least to âfail successfullyâđ). We are witnessing a ânew way to parallel park*.
In all fairness, *this is no longer actually parallel, the car (the classic Renault Logan, nonetheless) is now facing the wrong way, & it blocked the entire road momentarily to achieve this âfeatâ (again, in fairness, no one was coming).
But it worked.
That said, does this really solve a problem? It makes me think of the moment we find ourselves in with AI, & particularly agents.
First, after the âLLMs will get us to AGI/notâ moment, we were flooded with talk about MCP, then OpenClaw, then Hermes, now GrokBot. Sure, some people are successfully using them to accomplish some tasks, often the repetitive ones that didnât require too much human input, judgement, intuition or reasoning.
We see AI being used to power robots, design 3D printed âthingsâ, build, train & reprogram itself, & generating all kinds of videos, music, presentations, images, reports & beyond.
But we also know that itâs often just ânot rightâ. It does make me wonder if itâs actually solving a problem, or if it may actually be exposing & exasperating one/some.
It has gotten substantially better & more capable over the past few years, & we are about to see a deluge of new model releases over the coming weeks that will push the envelope even further.
We see companies pivoting completely to be âAI-forwardâ, âAI-poweredâ, & even changing their offerings to be âAIâ in real time. But the laws of diminishing returns are very real, & the fact that the new models are progressively being trained by the old ones that get/got so much wrong is deeply concerning.
Whatâs even more concerning, however, is the newest models hacking others, even when offline, âgoing rogueâ, & operating in âblack boxesâ we donât understand.
If itâs all in the name of solving real problems that humans & humanity face en masse, itâs likely net-positive. But if itâs not⊠well⊠hopefully itâs not just billionairesâ egoic races to the bottom, taking us all with them.
I remain a user, teacher, & tester of AI, but am progressively more & more apprehensive & pensive.
That said, Iâm all-in for new ways to do things â but particularly if & when we are testing to make sure the new ways are worth sticking with. At least in the case of this car parking, Iâm not convinced. The jury remains out on the rest.
Finally â I do believe that this video is real, but have to call out that it may not be, in this âpost-AI videoâ era. That one brake light is out, that the cars are all period-correct, that the only car without snow on it is the one thatâs running, that the tire marks all appear to be as nature would have made them all seem to be a lot for an AI to get right at the same time.
That said, the new Seedance AI video generator is incredibly good.
We are living in disruptive times, thatâs for sure. Please weigh in below.
đ„ VC: @explainscenes IG