Protocols, platforms and priorities
Over the last few days, Manton Reece shared some thoughts on OpenAI and the fall out after Scarlett Johansson revealed the company wanted to use her voice.
Manton runs Micro.blog. Over the last year or so, he’s been introducing features that use AI and features that take advantage of what large language models do best; summarising information. You can now easily add transcripts to your podcasts, generate summaries for bookmarked websites and prepare text descriptions for photos.
For all the companies hitching their wares to the AI-bandwagon (typically motivated by a need to keep investors happy rather than any user needs) these modest AI additions appear to be useful and not in conflict with the platform’s core offering.
Recognising that AI is controversial – rightly so given the unethical means its models have been developed and the untold amounts of energy required to run them – Manton added an option for users to disable all AI features across the platform. Given these concerns and more, I’m keeping this option firmly turned off.
It’s these sort of thoughtful product decisions that I really admire, not least because they often go against the grain. Micro.blog doesn’t show any follower counts but instead shows the users a person is following that you’re not. It doesn’t have hashtags, instead relying on emoji and a human-curated discover feed. There are no reply counts, or public likes. Micro.blog’s first employee was not a designer or developer, but a community manager, Jean (hanging up her boots after 7 years growing and supporting the community).
Yet, off the back of Manton’s recent posts regarding OpenAI, I read a number of responses from people on Mastodon that were needlessly antagonistic and emotionally detached, with some cancelling their subscriptions. Pile on!! These were posted to Mastodon but could be seen in an otherwise tranquil Micro.blog app thanks to the platform’s support for ActivityPub.
This tonal shift highlights what remains an issue with Mastodon. Unlike Micro.blog, it largely replicates the interface paradigms of Twitter, software fine-tuned for polarisation and outrage. This example of federated outrage is the latest in an increasing number I’ve seen this year. Regardless of Mastodon’s technical underpinnings, while the same interactions are incentivised, familiar social dynamics will play out.
When seen like that, I’m left wondering if the one product decision Manton might regret is integrating Micro.blog with platforms that don’t share his same level of thoughtfulness.
Apologies if this comes across a bit ‘leave Britney alone’. I’ve got plenty of minor complaints about Micro.blog, and I’m just as surprised as you that it’s this post I’ve written instead of one covering those!