Today’s insight is brought to you by Mat Velloso, on the difference between machine learning and AI:
”If it is written in Python, it's probably machine learning. If it is written in PowerPoint, it's probably AI.”
↪ Replying to @jopas
@jopas Wow, beautiful insights and reasoning, and lots of stuff for me dig to into. Thanks for sharing!
I remember when Zuckerberg wrote A Privacy-Focused Vision for Social Networking.... because it was only a few months ago. 😬😫😖 x.com/jason_kint/sta…
↪ Replying to @unknown
@sladner @movito @shannonmattern Wait I thought he finalized the English version... I can check with him tomorrow. Meanwhile, here's our interview with Jonas from 2012(!)
uxpodcast.com/episode-32-jam…
@sladner @movito @shannonmattern It's not finalized yet. Sorry. Don't mention the war 😬
↪ Replying to @unknown
@hugofroesUX @jopas @luiseduardodejo @SerRodSan @nickf @beantin Many excellent points here and some points I don’t agree entirely with. But that is normal for me.😁
I wrote some quick reflections and attached them.
Accessible text here: telegra.ph/Reflections-on…
↪ Replying to @grandeped
@grandeped @hypervisible "Please leave the room and report to the proctorlogist."
↪ Replying to @hypervisible
@hypervisible Gah, that word sounds like a medieval torture contraption...
"Sir, we're going to have you in the proctor if you do not confess to these crimes!"
I'm researching the Dewey Decimal system for a metadata presentation.
The topic "Homosexuality" was classified
- in 1932 under 132 (mental derangements) and 159.9 (abnormal psych.)
- in 1952 also 301.424 (study of sexes in society).
- in 1989 also 363.49 (social problems).
This matters because classification systems are used in machine learning. Machines do not make value judgements but humans do. There is a very real risk that AI can slow the progress of changing values by learning from data where prejudice persists.
Once again, the people at risk in what is often called improvement are the people who are always at risk, always set a aside and rarely included in designing the future. They are the ever bypassed stakeholders of said improvement.
Researching the Dewey Decimal system for a metadata presentation.
The topic "Homosexuality" was classified
- in 1932 under 132 (mental derangements) and 159.9 (abnormal psych.)
- in 1952 also 301.424 (study of sexes in society).
- in 1989 also 363.49 (social problems).
This matters because classification systems are used in machine learning. Machines do not make value judgements but humans do. There is a very real risk that AI can slow the progress of changing values by learning from data where prejudice persists.
Once again, the people at risk in what is often called improvement are the people who are always at risk, always set a aside and rarely included in designing the future. They are the bypassed stakeholders of said improvement.
↪ Replying to @plannero
@plannero @ninafogelberg @lisasallvin @digitalsamtal @TheConfMalmo @blixtgordon13 @shoshanazuboff @jonastics Tack Pär 😊 Läser ni min handbok Nina så kan jag gärna komma förbi bokcirkeln och svara på frågor, och kanske ställa några också :)
↪ Replying to @JesperBylund
@JesperBylund @kazarnowicz @beantin @charbeliam Agree, these ideas look very much in line with my thinking.
"Most people want to do good things." While likely true, I'm not sure most people take the time to figure out what "good" means for them.
Also not sure design should be the ethics department ;)
↪ Replying to @unknown
@radicallyrach @movito Getting along, building relationships, respecting each others work and being able to provide value seems to be the gist of it.
↪ Replying to @JesperBylund
@JesperBylund @letterpress_se Interesting. Do you know if they estimate the likelihood of the respondents having incentive enough to give a truthful (and representative) answer?
x.com/weylandswart/s…
But but I thought all computers are neutral and unbiased and perfectly just and error-free?
#digitalethics
theguardian.com/technology/201…
↪ Replying to @charliesaidthat
@charliesaidthat @danbarker Like the writing, not the subject matter.
↪ Replying to @danbarker
@danbarker @charliesaidthat I like this article that traces it to 2015. lifehacker.com/is-this-an-inn…
1. Tetris
2. Pitfall
3. Leisure Suit Larry in the Land of the Lounge Lizards (I was 13!)
4. Monster Truck Madness (playing online in late 90s)
Karateka, Lode Runner, Lemmings and Ultima III:Exodus deserve honorary mentions. x.com/MrMidasGames/s…
Actually, would be interesting to hear if anyone knows anyone with this role: read why people unsubscribe to newsletters and then act on any insights. x.com/axbom/status/1…