Two brief updates to membership:
- We changed billing systems the other week. Our thanks to those who signed up after the change and called out a bunch of old cobwebs that we forgot to clean up. If you were on the old system, we sent you an email that you have a couple of weeks to switch over; we’ll be turning that one off on August 01.
- We’re offering free half-hour calls to new members going forward. Easier than our previous hours, and more value for you.
Now would be a great time to join us as a member, especially if you work in software. We’ve got a lot of great lessons planned over the coming months.
I’m going to close this brief series of pieces on post-LLM design on some thoughts about how I approach any new technology.
I’m mostly motivated by Ursula Franklin’s principles in her phenomenal talks about technology, which you should read. In her first lecture, she said:
Any critique or assessment of the real world of technology should...involve serious questioning of the underlying structures of our (production) models, and through them, of our thoughts.
I think about these things all the time. All technology comes from an intention, and it has an impact. Someone stands to benefit from it. Others may be harmed by it. We don’t do enough to look at a technology and ask what its effects will be. We adopt it blindly, because others say we should, and let the rest of society figure it out. This renunciation of personal agency is deeply concerning to me, so I’ve spent my whole career trying to preserve it in myself.
Machine learning models have been around for many years, and you’ve absolutely used them in your daily life. Every time your spam filter traps something or you’re subject to a complicated algorithm that guesses whether you’re pre-approved for a mortgage or credit card, you’re witnessing some proto-LLM technology at work. Cheap computational power & larger data sets have made LLM more feasible over the years, so now here we are.
The ontology of LLM, the existence of LLM, can be reviewed separately from what we’re doing with it. We choose to use technology in specific ways. The iPhone came out; a couple of years later, we’re doomscrolling, destroying the taxi industry, and texting our friends at all hours. We chose to do these things.
And the things we’re choosing to do with LLM are broadly not great. I am, for the record, profoundly unhappy with people incorrectly thinking that LLM is a useful justification for mass layoffs – but I get it, because capital will find any possible way to automate labor out of existence. I’m also unhappy with the cultishness of most LLM companies. I’m unhappy with the climate impact of LLM. I’m unhappy with LLM ripping off the literal canon of civilization to turn out reams of mid-grade nonsense. I’m unhappy with people getting addicted to LLM. I’m unhappy with people using LLM as a crutch for real human connection. As a result, I don’t view LLM as a net positive development for society.
There are a few bright spots. Using LLM locally on my Mac Studio has occasionally been helpful. For blind & low-vision people, LLM-powered smart glasses have positively changed their lives. But in a broader sense, I look at what we’re doing with LLM, and I ask: gosh, why?
So I’ve been trying to write about post-LLM design as a way of centering the function of design in a changed world. LLM will not kill design, as we’ve discussed. LLM will create a lot of noise that may harm the short-term acceptance of design. At the end of the day, that may create a lot of upheaval in our industry that I think is not positive, and which is a conscious choice by those who buy design. Many smart, kind, talented people will leave our industry. That sucks, and everybody howling about how those people should have learned LLM is profoundly missing the point.
The goal, then, is to look at what is happening, do everything in your power to see it for what it is, take the parts that might work for you, and ignore the rest. You can also, contrary to prevailing wisdom, move slowly & intentionally, letting the dust settle before you act. I’ve done all of this, having barely written anything about LLM prior to this series. I regret nothing.
The issues surrounding LLM have become so tangled up in the actual technology that to even talk about “LLM” automatically thrusts us into this horrible spiral of what-ifs about what it will do to all of society. Put another way, the technology itself is different from the labor issues that those in power are purposefully creating as a response to it. Knowing that, then, how will you move?