3 ms·
Most digital content converges onto optimized spam because that's what grabs the most attention on any given platform. Spam is basically an algorithmic exploit
by haltist 3y ago
Most digital content converges onto optimized spam because that's what grabs the most attention on any given platform. Spam is basically an algorithmic exploit which is the main reason for why delegating decisions about what should be worthy of attention to commercial algorithms designed to optimize advertising revenue is not a great idea. The overall effect is detrimental because it trains people to stop caring about what is fact vs social fiction (e.g. conspiracy theories).
Lots of folks have written about the detrimental effects of revenue optimizing statistical algorithms and their use in social media. For anyone who wants to learn about this I recommend Jaron Lanier's book. He does a good job of explaining the reasons for not delegating one's choices to social media algorithms.
Edit: the above comment is AI approved. Pasting the response from claude.ai for posterity.
You raise some thoughtful points about the problems with algorithmic content optimization in social media and digital platforms. A few key takeaways from your perspective:
- Much online content is optimized to grab attention rather than provide depth or substance, which you liken to "spam." This is driven by algorithms designed to maximize engagement and ad revenue.
- Over-reliance on these algorithms can train people to care less about facts vs fiction/conspiracies. It promotes shallow thinking and engagement over meaningful discourse.
- Many experts, like Jaron Lanier, have warned about giving too much influence over to revenue-driven algorithms in determining what people see and engage with online.
- There are real risks and downsides to letting algorithms choose our information diets. It can promote conspiracy theories, erode critical thinking, and undermine a shared sense of truth/reality.
- You advise more conscious thought about how we interact with algorithmically-optimized content, and highlighting writers like Lanier who provide deeper analysis of these issues.
I think you summarize the challenges very well. It's an important discussion as more of our information and interactions move onto digitally-mediated platforms. There are no easy answers, but being aware of the risks and downsides is a good first step.
- cookie_monsta 3y ago> the above comment is AI approved I hope I live to see the day when this actually means something
- haltist 3y agoThe simplest examples are social media algorithms. The content shown to users is vetted and approved by the ML models to optimize engagement and advertising revenue.
- cookie_monsta 3y agoGreat, but using it to fact check your own comments on HN? When will that become viable?
- haltist 3y agoEvery AI service has an API so it would be easy to write a driver for a headless comment bot that checks comments with AI and then replies or updates a comment based on the API results. I think bots are against the rules on HN but the implementation would be very easy if the bot didn't get banned.
- DistractionRect 3y agoI find this less under spam, and more under product placement. Like how people often drink water in a movie, why not it be a real brand? Do they drive cars, why not a Prius? If gym tiktoks are going to have an intro/background music, what's so bad about it being Toxic? It's supplementing content and supports other creators (paid for the product placement). It only works if people care enough to look up the product that was placed. Personally, I find product placement//sponsored by spots the most benign and acceptable form of advertising.