OMG — I hate acronyms and abbreviations ;), what is t2i and i2i?
t2i - text to image (text prompting, prompt engineering)
i2i - image to image (includes use of controlnets and ipadapter and upscaling)
1 Like
Whether media is creating the GenZ work ethic (or lack thereof) story or simply reporting on it, that does seem to be a trend, likely because so many of them came into a cash-bloated economy that has suddenly slammed the brakes and door on their careers.
Can vouch for ‘GenAlpha’ missing two years of education. Its depressing but the new normal for the US.
1 Like
@ MK19 a couple updates please?
just before school started you mentioned your students using simple AI tools to demo an ID workflow, have those tools/process evolved any?
and you mentioned that soon, students would arrive more fluent in AI than the professors is that still going to be the case?
1 Like
It’s interesting to return to this thread a couple of years later. I’ve been teaching for the last 5 years pretty steadily. One thing I notice is that most of the better students sketch or want to learn how to sketch.
In my limited experience it seems that the students who default to AI tend to also be not as good in other areas.
It’s almost like a lack of wanting to put the work into learning and perfecting something is a tell to the student’s overall attitude to the subject matter.
My classes don’t have grades so I always tell the students in the end they can do what they want, but if they want my feedback then they have to put in work. If they don’t my then why should any of us in the class spend time giving them critique?
But also, every action has an equal and opposite reaction. I’ve noticed that the more the tools are pushed on them from all corners of the internet the more they tend to not want to use them and learn more durable skills. Especially in the last year. I always tell them to not make their career dependent on a cloud based platform that they don’t own or control and that can be restricted at any moment.
You learn the theories by doing. Then you can learn shortcuts and cheat codes once you have a firm grasp of the fundamentals.
Friday thoughts.
2 Likes
I think AI’s a good news/bad news situation for design schools. I agree with Bill Gates and Karim Rashid recent statements, but it’s too soon to panic. Tech layoffs caused an evaporation of entry level jobs in coding, so freshman enrollment in computer science this fall has dropped by over 8%. Just about the safest career path is now the harbinger for all white-collar majors. That said, Ford just re-hired hundreds of senior engineers to fix AI mistakes, so the ability to move up the skill chain has been overstated.
Karim points out that AI’s best at rules-based tasks, which is true. Therefore, an ID major will be seen as a safer bet for a career because of the “fuzzy front end” and the fact that human nature is the most complex and least understood component of any system we design. Enrollments should increase for years to come. That won’t last forever, Karim thinks schools are already producing a glut of designers today. One example of what to teach might be the how/why of using VBL. Creating the guide is complex compared to executing the iterations (tasks done by a designer vs interns or, eventually AI) Once entry level jobs start to dry up design education will have to adapt.
What that change looks like may look like some of the experiments that have taken place. One struggling department in a midwest art school (with a 50% placement rate) partnered with a local design research consultancy to reinvent their curricula towards human centered design. Craig Vogel started an industry/academic innovation hub at DAAP and Don Norman’s program at UCSD show re-invented design as innovation at the undergrad and graduate levels.
I believe that if/when entry levels do dry up the schools that stay in business will have robust internship programs. Ultimately, if it’s at all possible for a student to leapfrog into a mid-career position, engineering and business schools will have to figure it out first. So we can learn from that. But as Bill (and Geoffrey Hinton) predict, saving jobs will probably it’ll take some government regulations to ensure.
2 Likes
Why would anyone care about what Karim Rashid thinks? Your opinions carry just as much or more weight.
@no_spec I’ve long believed this has been why UC grads have come out of school in such a good position. I worked with a UC grad right out of school and then several at Nike. Because of the internship co-op program they basically graduate with 2 years of experience. And they all benefit from sharing their experiences when they are in the academic cycle of their year. It just seemed like they had a firmer grasp of what they wanted to do and what they were good at having 4 co-ops before graduating.
2 Likes
Someone here once said there’s a difference between schools that prepare you for your first job and those that prepare you for your last job. The differing agenda is the cause of the debates around fitting ten pounds in a five-pound sac
the Co-Op/internship systems argument is that getting people into jobs is the first priority and they can learn everything else from mentors or YouTube. It seems to work out that way for UC.
I mentioned Karim’s comments because he’s representative of another perspective on education. All those architects and designers who were inspired by, and aspire to have work in museum collections are best served by the Bauhaus approach. He said designers in the age of AI will have to work harder to be successful. By which he means visual novelty. He definitely understands that an AI could be trained on his body of work to reproduce his signature style (a rules-based VBL) without him lifting a finger ever again. To remain relevant, he needs to re-invent his entire aesthetic, and if he was smart he’d let AI take over the day to day while that became his main task. His new visual brand language would have to be avant guard, and of course eventually re-reinvented as it became more mainstream. This is not a workaday approach to ID, more fine arts
I don’t know which direction AI will push design and design education but I think we’ll have more time to figure that out than other white-collar professions.
I need to disagree there. I think there is zero benefit for him to use AI, and in fact I think it is better for designers like this who lean into a more artisanal or collector realm to clearly state they use no AI in their work.
I wouldn’t read a book made by AI no matter if it was trained on my favorite author’s style. Just zero interest.
Think maybe you guys are saying the same thing… Kreemy can continue to print pink money having the bots come up with formulaic regurg, while he points his considerable intellect to the ‘artisanal collector realm’. Not a bad gig if you can get it.
oh no, I’m definitely clearly saying he should lean into a “no AI” stance and not do the formulaic stuff. Any use of AI with his work or by anyone else in that space essentially devalues all of it to zero in my opinion.
Interesting to hear what Rashid thinks. Thoughts are needed, it is a pivotal time. Less interesting are driveby conversation dismissals.
Also disagree, making this a known action would kill his reputation. However, if any work is formulaic the market assumption will be that it is slop. Hence the re-invention pressure.
1 Like
perhaps “take over” was a mistaken way of saying that he should shift his emphasis towards inventing the next version of himself vs overseeing stuff going to Target. I’m thinking of how the fashion industry has already got knockoffs in production before it shows the haute couture in Paris. The ability to stay relevant will be staying ahead of AI’s capability to discern a pattern well enough to make derivations. that’s all I’m suggesting there
1 Like
it might help to talk about slop a bit, just to define it and it’s impact on education as noted earlier (letting ai do the work instead of thinking through the design)
The problem stems from the ability to create very finished seeming results without any training or experience. Particularly problematic where the barriers to entry are low. A few sentences and out pops a book, illustration, song etc. Anyone can stand on the shoulders of a world of genius without the discipline of having learned the craft. With little to no cost or personal investment or responsibility for the results. So, with such low stakes the bar is low by default.
Raising the stakes started a few years ago with the actors+writers strike in Hollywood, extrinsically forcing out AI competition. But it’s also possible to raise the stakes intrinsically elevating the bar, which has happened on at least one case. Last year the #1 country billboard chart song “Walk My Way” was created by a kid in his bedroom after months of working with AI. Could’ve settled for crap but pushed himself for something more. The question is: does anyone still need to learn to sing and play guitar to succeed? (in Country Music it’s definitely yes because of ‘authenticity’ being central to the brand, but you get my point)
Now to ID. Sexy renderings have always appealed to students as a workaround for a thoroughly thought-out design and nothing makes that easier than AI. The challenge for educators then as now is to determine if the student is not getting it, not capable of seeing the big picture or too intimidated or just not that into design. They’ve got to care enough about the entire job description and the entire process.
AI is not a solution but a tool, and because the variety of expertise plus significant finances needed to put any mass market product on shelves means the stakes are and will remain too high to tolerate slop.
1 Like
“AI is not a solution but a tool”
Someone once said about the U.S., “America is so big, everything you say about it is true.”
The same holds for AI.
AI is an umbrella so big that everything you say about it is true. Tool, crutch, collaborator, stupefacient, grift, theft machinery, fifth horseman of the apocalypse, magical. Take your pick. More than any other previous system, the knowledge and/or inclination you bring to the table determines the character of its contribution.
My exposure to design educators is on LinkedIn. The takeaway, they are also unsure about the future of design education. They seem to be as much at the focal point of job uncertainty as anyone else, overly embracing AI systems, seemingly to stay relevant. If reliance is on educators to individually figure things out in terms of AI for students, a good overall outcome can’t be expected.
1 Like
As an update: I have been teaching AI in many areas for a few years now. Industry advisors and external examiners (also in industry) are very impressed with it all.
Most academics are not practitioners, only 1/10 people that teach industrial design in the UK can teach anything practical (be that sketching or anything digital or CAD related), so it’s not surprising that students also pick up AI faster than them and can use it to greater effect.
While personally being generally pro AI, and facing resistance from career academics that have never designed and cannot do so today even using industry standard tools, when push came to shove recently where students are submitting purely AI generated works it turned out I was the most strict on it and trying to get it stopped. For academics whose only interest is publishing rambling pseudo-philosophical opinion pieces and design theory grounded in no practice, teaching is anti-thetical to their career development. So when teaching becomes easy for them (just pass the slop because sadly the slop is better than half student-made work now) they were happy to accept it en masse.
The gap is only widening. I hope young people thinking “AI can just do it, no need to get a degree in it” reduces the number of people studying design overall and we are left with more of the cream of the crop. Sadly, I foresee quite a different future. Bright students will see a trajectory where employment is more sparse and salaries shrink and choose to study mechanical engineering, or even better, robotics, and we will be left with weaker average students. Lots of people that think design is about self expression (ala Karim Rashid, Ross Lovegrove, etc.) and AI can help them realise their dreams easier and that their special snowflake mental creations are the sole value in designing things.
2 Likes