There could be something very interesting brewing from the folks behind Markforged.
The two individuals involved are Greg Mark and David Benhaim. If donāt recognize them, you should, as they were the force behind Markforged, now one of the leading companies in 3D printing.
Mark, whom Markforged is named after, was the first to develop continuous carbon fiber 3D printing. Benhaim was the co-founder of Markforged and was its CTO for many years.
Both are no longer with Markforged. Mark eventually shifted into board chair, but then departed at the end of 2021. Benhaim departed Markforged in July 2022. No doubt both benefited greatly from Markforgedās entry to the stock market in mid-2021.
What have they been doing since then? Where have they been spending the cash from Markforged? Thatās been a bit of a mystery, at least until now.
Correction: it is STILL a mystery, as you will see.
Mark posted a mysterious message on LinkedIn this week, suggesting something ābigā is in the works. There were basically no details, and the only image provided was the āpresentā shown at top.
However, we can do a bit of espionage here and try to guess whatās happening.
Mark said:
āDavid Benhaim and I are back at it. Buckle up.ā
Possible Implication: They and a small team are building a new type of machine, process, or software, very likely related to additive manufacturing.
Mark commented:
āThe garage is still my favorite room. Technically, it’s a hanger now.ā
Possible Implication: Aerospace applications?
Markās LinkedIn profile said:
āInvented carbon fiber 3D printing (NYSE:MKFG). Now we’re building the big one.ā
Possible Implication: Large-scale device? Scalable manufacturing platform? Extended use of continuous carbon fiber?
Both Mark and Benhaim list āStealth Startup – Generative AIā as their current full time gig, with Mark being Founder and CEO, and Benhaim being CTO.
Possible Implication: The new project could be a software play leveraging generative design approaches, not a hardware device.
My suspicion is that this could be an application of modern AI tech against generative design tools. Current generative design tools are provided a list of constraints, and then they iteratively āgenerateā the minimum required structure to handle those constraints.
Using an AI approach for a similar purpose could be quite interesting and provide some advantages. One could be efficiency: current generative solutions are currently pretty heavy. Meanwhile, AI is heavy in one-time training, but not so much in repeated execution afterwards.
Iāve speculated on some uses of AI in 3D design previously, and weāve seen some of this already emerge with a couple of ātext to 3Dā solutions, including OpenAI’s Point-E and Luma AI. However, Iāve always thought that this concept could be better implemented for industry by training on a large number of existing part designs. Modern AI systems use a large set of input data for training. I am not aware that anyone has done so using CAD models. Could their startup be a āText to Partā system? Maybe it includes not only the generation of the 3D model, but also its production?
We cannot yet know what exactly the pair are up to, as they donāt even provide a name for their new company. Instead theyāre using a generic āStealth Startup – Generative AIā as the placeholder.
Perhaps the real company name would give away what they are doing.
Regardless, this is absolutely something to watch: two very well funded and highly experienced individuals in the 3D space working on something big.
Via LinkedIn