“I fed a picture of donald duck to my computer and asked it to identify said duck. I then asked my computer to draw donald duck’s girlfriend in svg format on my shitty decade-old gtx 1080. It ran into an infinite loop since the number of polka dots kept exceeding its context window. So finally I decided to generate a textured 3D mesh from donald duck and it did so within 5 minutes without much error, then I approved for my computer to send it to my 3D printer”
This comment was not something I would have made 10 years ago, nor would have expected 5 years ago, even 3 years ago once LLM quantizatization largely enabled LLMs to be run locally on 8 gb of vram. Nor did I think it would have been a comment that I would have made 6 months ago. It’s not exactly surprising, except for how quickly I was dragged into it, within the past few months, with my limited resources.
If we split the comment into sentences associated with years that it became plausible for an enthusiast with unspecialized hardware:
“… decade-old gtx 1080 – 2016
“… asked my computer to identify said duck”1 – 2021
“… draw donald duck’s girlfriend in svg format1” – 2022
“… on my decade-old gtx 1080” – ~2023 - 2024
“… generate a textured 3D mesh from donald duck” – ~2023
“… within 5 minutes” – 2025
Try saying that in 2006, a decade before the gtx 1080 was released, just as PCs were tipping from 32-bit hardware to 64-bit.
We keep declaring computational niches dead, and move on to the next new thing. With significant constraints on hardware availability, there is now more human attention towards revisiting these niches to make programs more resourceful. By then we were already reviving decade-old hardware, extending their capabilities beyond what was imagined at their conception.
Even with the machines that we make, we don’t know the full extent of their capabilities. And probably can’t. As Donald Knuth puts it, “many more algorithms exist than anybody can understand or ever make use of, or discover, because they’re just way beyond human comprehension. The total number of algorithms is More than mind-boggling. So, we have situations now where we know that algorithms exist, but we don’t have the foggiest idea what the algorithms are.”
At some point, we found a way to find algorithms that we couldn’t understand on hardware that we do not know. LLMs have passed the threshold of human energy and comprehension for certain kinds of problems. These are algorithms that have found counterexamples to the Jacobian conjecture, the Navier–Stokes smoothness problem, and possibly Exponential Time Hypothesis – the latter mostly without human assistance – which might overturn decades of ETH-based complexity problems.
This is the kind of scenario that Hans Moravec wrote about in Intelligent Machines almost 50 years ago. “The expanded abilities of existing systems such as Macsyma, the symbolic mathematics system from MIT, which can be used as a desk calculator for doing algebra and trigonometry as well as arithmetic, along with new experimental results, would accelerate theoretical development. Gains made this way would improve the very systems being used, causing more speedup. The intermediate results would be inefficient kludges busily contributing to their own improvement. The end result is systems as efficient and clever as any designed by more theoretical approaches, but sooner, because more of the labor has been done by machines.”
Much of Moravec’s writing considers machines as part of an evolving ecology. That is, systems morphing and shifting in a physical space – consisting of proteins, biological organisms, topographical shifts and weather patterns – with boundaries between them that give rise to their increasing variety and resourcefulness. Algorithms will find the same dynamism in memory space, where bit-length and endianness, userland and kerneland, virtual memory and NIC DMA zones are mere mountains, valleys, rivers and caves of a living terrain. Electromagnetic interference and voltage line noise mere weather patterns. Evolving, inventing, constructing – on seemingly arbitrary bit-lengths woven like synthetic threads on a sheet. I wonder how far it will go.