I always considering completely escaping the consumption of algorithmic based content when I read your newsletters but then I realize that I woukd have probably never found you if I didnt scroll. Hard decisions to make. So many central aspects of my personality and things I know about are thanks to the FYPs. But there is also an enormous negative side such as the aspects you discuss or simply doomscroll. Btw attention span isnt something that is being damaged in my opinion. This is a widely pushed fact and that people are able to learn less effectively because of this. I think this is not true because consuming fast paced informations has allowed me to learn at a faster pace but in a different style of course. Once I realized this I simply had to break free of conventional teaching methods and find ways to engage myself with enough stimuli and at the right (faster) pace. The results has been astonishing, studying isnt something boring and slow anymore and I always look towards it like a long session of a challenging video game. I think this should be discussed more. The educational system is extremely outdated and needs a reform.
Some of my favourite songs have come from the algorithm, many graphic design tools I've found came from the algorithm, it's a fantastic resource for creatives. However, I've also wasted countless hours on there. How does one weigh up the positives and the negatives?
I try to see algorithmic based content as entertainment first, not as legitimate political or cultural media, with the occasional gem spread throughout. This content is here to stay, and I think as long as we don't confuse the role it plays, then we can navigate this.
This is a great summary, thought I’d share some relevant facts:
- In data science/engineering we use the following data classifications: Missing Completely at Random (MCAR), Missing at Random (MAR), or Missing Not at Random (MNAR). The plane example would be considered MAR. Definitions for these are easy to find, and I highly recommend familiarizing with this concept since classification itself is its own world of controversy (this would be a fantastic post in itself).
- The “algorithm” is actually a two part problem: defining “success” is a separate process from defining the algorithm itself. This success criteria is often Goodharted, leading to the notorious negative feedback loop you often see today. This isn’t a necessary outcome in my opinion, but fixing it requires nuance that’s hard to measure and justify.
- Personalization can extend beyond what you engage with, and potentially encompass your personal engagement style.
One thing I would like to add is that it creates a feedback loop of 'negative' content, people engage with it -> they get polarized over time and create more negative content.
Understanding algorithms gaze makes me continuously less interested in using other social media (yes, Substack is still kinda social media, albeit my fav)
The idea of filters have always been there: editors of newspapers, anchors, and what happened to Janet Jackson and Usher at that Super Bowl?
I’d be naive to think not being on social media as much as I can removes all filters from how I interact with the world, but that’s never been true.
The internet and social media are both great sources of information, but what people are forgetting is that they absolutely can not be your only source of information.
As you've said, the path that information takes to get to you is heavily modified. And because of that, it's only going to represent a portion of the conversation. You've got to step out of the digital space to see the broader information landscape.
During weekend around, Instagram filter for gore went off. I wasn't on it, but there were videos and mentions from others on it.
It is kinda weird to realise what we got only when it is gone, in this aspect/allegory!
Makes me wonder, if algo weights fell down, what would we see? Nothing, or would we see all posts? Would fb/ig/yt fart itself into DNS error from overblow?...
im so grateful people smarter than me are discussing the algorithm and it’s impact on politics. Really good read, thank you :3
I always considering completely escaping the consumption of algorithmic based content when I read your newsletters but then I realize that I woukd have probably never found you if I didnt scroll. Hard decisions to make. So many central aspects of my personality and things I know about are thanks to the FYPs. But there is also an enormous negative side such as the aspects you discuss or simply doomscroll. Btw attention span isnt something that is being damaged in my opinion. This is a widely pushed fact and that people are able to learn less effectively because of this. I think this is not true because consuming fast paced informations has allowed me to learn at a faster pace but in a different style of course. Once I realized this I simply had to break free of conventional teaching methods and find ways to engage myself with enough stimuli and at the right (faster) pace. The results has been astonishing, studying isnt something boring and slow anymore and I always look towards it like a long session of a challenging video game. I think this should be discussed more. The educational system is extremely outdated and needs a reform.
I've found myself in a similar situation.
Some of my favourite songs have come from the algorithm, many graphic design tools I've found came from the algorithm, it's a fantastic resource for creatives. However, I've also wasted countless hours on there. How does one weigh up the positives and the negatives?
I try to see algorithmic based content as entertainment first, not as legitimate political or cultural media, with the occasional gem spread throughout. This content is here to stay, and I think as long as we don't confuse the role it plays, then we can navigate this.
This is a great summary, thought I’d share some relevant facts:
- In data science/engineering we use the following data classifications: Missing Completely at Random (MCAR), Missing at Random (MAR), or Missing Not at Random (MNAR). The plane example would be considered MAR. Definitions for these are easy to find, and I highly recommend familiarizing with this concept since classification itself is its own world of controversy (this would be a fantastic post in itself).
- The “algorithm” is actually a two part problem: defining “success” is a separate process from defining the algorithm itself. This success criteria is often Goodharted, leading to the notorious negative feedback loop you often see today. This isn’t a necessary outcome in my opinion, but fixing it requires nuance that’s hard to measure and justify.
- Personalization can extend beyond what you engage with, and potentially encompass your personal engagement style.
hey, the diagrams with transparent backgrounds are nearly unreadable when in dark mode. Great stuff otherwise !
thank you, updated!
One thing I would like to add is that it creates a feedback loop of 'negative' content, people engage with it -> they get polarized over time and create more negative content.
This is beautifully written. Thank you for using examples to simplify these concepts.
Understanding algorithms gaze makes me continuously less interested in using other social media (yes, Substack is still kinda social media, albeit my fav)
The idea of filters have always been there: editors of newspapers, anchors, and what happened to Janet Jackson and Usher at that Super Bowl?
I’d be naive to think not being on social media as much as I can removes all filters from how I interact with the world, but that’s never been true.
YOUR NAME IS ADAM???
The internet and social media are both great sources of information, but what people are forgetting is that they absolutely can not be your only source of information.
As you've said, the path that information takes to get to you is heavily modified. And because of that, it's only going to represent a portion of the conversation. You've got to step out of the digital space to see the broader information landscape.
Fantastic
Great work! Love how you interpret the effects of algorithms in the news media we interact with. Loved it
During weekend around, Instagram filter for gore went off. I wasn't on it, but there were videos and mentions from others on it.
It is kinda weird to realise what we got only when it is gone, in this aspect/allegory!
Makes me wonder, if algo weights fell down, what would we see? Nothing, or would we see all posts? Would fb/ig/yt fart itself into DNS error from overblow?...