What do you think?
Steve: A Framework for AI and Identity Design
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 3 of 3 reviews
December 6, 2024
the first chapters, the insights into the design processes of other design studios, their comments and thoughts about identity design, are interesting to read.
but when it gets more detailed and as soon as the author draws her first big conclusion, based on the interviews and observations, she took the wrong direction, imo. why focusing on feedback?
if existing models are not yet fully adapted to designers’ needs and if they have to be trained extensively, why not ask for peer feedback in the studio? what is the benefit of the ML in this case? how does it improve the already existing feedback methods? this doesn’t add any new aspect. not to mention the unsolved issue of training the ML and which data set to use to train it with.
it is inexplicable why she focuses her research on this feedback idea. what the design studios mentioned and wished for on the first pages were either helping with getting an idea to paper, to avoid repetitive and tedious tasks, to facilitate visual research or to assist with the communucation between the designer and the client. nothing of this will be addresses with steve.
appearence of the developers: it feels as if they take the whole framework apart. okay, little design researcher with your cute idea, but now let me explain you something. it is not so easy as you think.
even after the revision of the framework, where she made some useful adjustments, like including some kind of visual output, the initial problem (focusing on feedback) remains.
on top: the examples that are provided are useless. the simplification as an illustrative purpose doesn’t help at all and reveals all the flaws: round = soft, sharp = agressive. so, round is better for children, wow. that is so biased and only reproduces common misconceptions of design stereotypes. it is one step before pink is better for girls and blue for boys.
also the second example. who wants to see, based on ones own sketches, what others already did? in fact that is the most demoralizing thing that can happen. if the concept or idea already exists it is immediately dead. and if not, this promotes exactely what she herself warned for a few pages earlier. that everything will start to look similar based on some global design trends.
why not use generative models to inspire ideas, instead of giving verbal feedback or referencing real examples? yes, they are also based on work that has already been done. but at least it is already kind of a remix?
if the book would just have explained in detail the individual processes of each design studio, that would have been so greatly fruitful. what our industry lacks is a broad exchange of processes. to read about pentagram’s and studio dumbar’s technical approaches was eye-opening. please explain exactely this, for all of us. we would have learned so much and probably a lot more than from an assistant called steve.
but when it gets more detailed and as soon as the author draws her first big conclusion, based on the interviews and observations, she took the wrong direction, imo. why focusing on feedback?
if existing models are not yet fully adapted to designers’ needs and if they have to be trained extensively, why not ask for peer feedback in the studio? what is the benefit of the ML in this case? how does it improve the already existing feedback methods? this doesn’t add any new aspect. not to mention the unsolved issue of training the ML and which data set to use to train it with.
it is inexplicable why she focuses her research on this feedback idea. what the design studios mentioned and wished for on the first pages were either helping with getting an idea to paper, to avoid repetitive and tedious tasks, to facilitate visual research or to assist with the communucation between the designer and the client. nothing of this will be addresses with steve.
appearence of the developers: it feels as if they take the whole framework apart. okay, little design researcher with your cute idea, but now let me explain you something. it is not so easy as you think.
even after the revision of the framework, where she made some useful adjustments, like including some kind of visual output, the initial problem (focusing on feedback) remains.
on top: the examples that are provided are useless. the simplification as an illustrative purpose doesn’t help at all and reveals all the flaws: round = soft, sharp = agressive. so, round is better for children, wow. that is so biased and only reproduces common misconceptions of design stereotypes. it is one step before pink is better for girls and blue for boys.
also the second example. who wants to see, based on ones own sketches, what others already did? in fact that is the most demoralizing thing that can happen. if the concept or idea already exists it is immediately dead. and if not, this promotes exactely what she herself warned for a few pages earlier. that everything will start to look similar based on some global design trends.
why not use generative models to inspire ideas, instead of giving verbal feedback or referencing real examples? yes, they are also based on work that has already been done. but at least it is already kind of a remix?
if the book would just have explained in detail the individual processes of each design studio, that would have been so greatly fruitful. what our industry lacks is a broad exchange of processes. to read about pentagram’s and studio dumbar’s technical approaches was eye-opening. please explain exactely this, for all of us. we would have learned so much and probably a lot more than from an assistant called steve.
Did Not Finish
August 25, 2025I guess I don’t care enough about AI or identity design enough these days. Hopefully the author got her Ph.D, though.
August 19, 2026
The back cover promises support for designers looking for practical advice, ethical frameworks and more. Unfortunately, the book doesn't contain a single practical guide on how to actually apply AI in the creative process.
This could have been a blog article. Most of the content is academic filler where the author describes what they did or didn't do in their research, rather than sharing findings a working designer could use. The research itself feels very subjective and is built on a fairly thin hypothesis. In practice, it comes down to interviews with a few dozen studios that don't really represent the market, they're all completely different from each other, which makes it hard to draw any general conclusions.
The interview questions also feel surface-level. With access to these people, I would have expected the author to dig into how they actually use AI tools in their day-to-day process. Instead, it reads more like polite small talk.
The main output is built on these subjective impressions: the author picks "feedback" as the area of the design process where AI should be applied, but it stays abstract.
To be fair, the book is beautifully designed,that has to be acknowledged. But rhere's a real disproportion between how the book presents itself (which is why I bought it) and the content it actually delivers. Beautiful form, but no substance.
My honest suggestion for the author would be to commit to one direction, either a design object, or genuinely meaningful research. Right now the book creates big expectations and then leaves the reader with wasted attention, and wasted money too.
This could have been a blog article. Most of the content is academic filler where the author describes what they did or didn't do in their research, rather than sharing findings a working designer could use. The research itself feels very subjective and is built on a fairly thin hypothesis. In practice, it comes down to interviews with a few dozen studios that don't really represent the market, they're all completely different from each other, which makes it hard to draw any general conclusions.
The interview questions also feel surface-level. With access to these people, I would have expected the author to dig into how they actually use AI tools in their day-to-day process. Instead, it reads more like polite small talk.
The main output is built on these subjective impressions: the author picks "feedback" as the area of the design process where AI should be applied, but it stays abstract.
To be fair, the book is beautifully designed,that has to be acknowledged. But rhere's a real disproportion between how the book presents itself (which is why I bought it) and the content it actually delivers. Beautiful form, but no substance.
My honest suggestion for the author would be to commit to one direction, either a design object, or genuinely meaningful research. Right now the book creates big expectations and then leaves the reader with wasted attention, and wasted money too.
Displaying 1 - 3 of 3 reviews