February 4, 2025

Deep research and open source

The arrival of DeepSeek has thrown a spanner into the old-fashioned mechanics of the US-China technology battle. Even incomplete verification of founder Liang Wenfeng’s narrative demonstrates how creativity plus limited resources can spur major advances, and at a fraction of expected costs.

America’s trade issue was once with Japan as Japanese industry, supported by its large protected domestic market, gobbled up the world of electronics, automotives and other things in the 1970s and 1980s. A major moment came when it switched from producing low quality cheap copies to the “copy and improve” stage and soon after to real standalone innovation.

Open source 

DeepSeek has given itself a boost by going open source, so cheap and easy for developers around the world to work with. This is a reminder of the influential 1997 essay “The Cathedral and the Bazaar” which looked at Linux and Microsoft and the advantages of open, lightly controlled forms of development over the traditional closed and secretive format. However, viewed collaboration in research is needed more than ever, so many clever new ideas come from mixing a number of old ones in a new way with one completely innovative element added to the mix; best uncovered where members of a wide network interact, be it formally or informally.

We have seen this in leather as in 1700s and 1800s sponsoring bodies like the Royal Society of Arts in London only supported scientists who agreed not to patent or protect their discoveries. The work of the 1850s Swedish chemist and inventor Carl Hyltén-Cavallius (Cavallin) on chrome tanning would have been properly utilised, and he be far better known, if he had not been trying to get to London to register a patent (he died enroute and only his obscure Swedish patent exists so he is largely overlooked in the literature). By comparison Dongola was a vegetable and alum tannage developed in Gloversville, New York, in the 1860’s by James Kent with the financial support of Booth & Co. to compete with the expensive glacé kid tannage for gloves. No attempt was made to keep the method secret. It failed as a glove leather process but was found to be excellent on lightweight footwear leathers from kidskins and kangaroo, finding a huge market in city footwear in rapidly urbanising America. Demand helped make many tanneries in Gloversville prosperous for some years and the Booth tannery was greatly expanded, with James Kent becoming very successful. The process helped the leather industry find a viable light weight leather and move towards the concept of “fatliquoring” rather than either using waxes and oils or ridiculously large numbers of egg yolks.

It is also fair to say that while in the chemical production side of leather making patents appear to be enforceable, it is much harder when it comes to leather production patents. At the same time over the past decade hiding behind pending patents and trade secret statements while making grandiose promises to the market has done biomaterials no good whatsoever. At the end of the day many were uncovered as not much improvement on a polyurethane soup with a little cellulosic content, with no indication of energy or water consumption in manufacture.

The pendulum swings east

What we should not ignore is that as the west declines in importance and the economic pendulum swings east we have major leather industries in India and China, as well as elsewhere in Asia and the Global South. Their undergraduate and then graduate students used to spend time in the west doing doctorate and post doctorate research. With the west turning against immigration this has become less attractive and with the growing economic dominance, demonstrated by increased trade, investment and economic activity many are finding local schools and institutes offer a better solution.

For some years Chinese researchers have had academic papers published in the UK JSLTC journal, and the quality of these is steadily improving. Like the American leather industry JALCA journal it is far from the being the best place to publish since as far as I am aware neither publication is accessible on-line through the University systems so authors and researchers are dependent on physical copies being present in their institutes and kept up to date. This means citations will be fewer and the JSLTC going fully digital will only make this worse. Leather papers are increasingly being published in other journals with higher ratings and a greater chance of being referenced, including new open-source vehicles.

Both China and India have some top institutes to build on such as Sichuan University in Chengdu and the CLRI (along with Anna University) in Chennai, and I was astonished last year during the 2024 Real Leather. Stay Different African judging to learn of so many bright students from leather and design schools which I had not previously heard of.

Papers from these top leather colleges and universities in China, followed by India and elsewhere are catching up and likely soon to go past the old countries. We have seen some powerful new offerings from western chemical companies but still lack a proper a replacement for high quality wet-blue, and the whole way we use solutions and suspensions of chemicals in water feels very outdated after so many thousands of years.

The leather industry needs some technological breakthroughs. Might we see something similar to DeepSeek happening in Asia for leather?

Michael Redwood

Leather chemist, writer, and advisor on responsible leather manufacturing and material strategy. This article was originally written for ILM.