Reddit’s 50% Plunge Fails to Entice Dip Buyers as Growth Slows
finance.yahoo.com
(Bloomberg) -- The gloomy sentiment around Reddit Inc. has failed to dissipate after its shares fell 50% from a February high, with volatile technology stocks under pressure. Most Read from BloombergGold-Rush Fever Returns to Historic New Zealand Mining TownWhat Frank Lloyd Wright Learned From the DesertBank Regulators Fight for Desks as OCC Returns to New York TowerThese US Bridges Face High Risk of Catastrophic Ship StrikesCharter Schools, Colleges Push Muni Debt Distress Near RecordThe social

The gloomy sentiment around Reddit Inc. has failed to dissipate after its shares fell 50% from a February high, with volatile technology stocks under pressure.

  • @[email protected]English
    0•21 days ago

    It’s not even about banning people, it’s about the fact that Reddit was never a sustainable business model from the start, at least not in the traditional capitalist sense where you’re actively trying to make a profit to please shareholders

    If they’d been stalwart about banning automation and keeping original, legit human content pure, they probably could have used it as a fountain of fresh data for AI, for polling, for engagement farming, and for promotion.

    The site was still growing even despite the admin induced atrophy. But they just couldn’t resist killing the Golden Goose.

    • @[email protected]English
      0•21 days ago

      I disagree that Reddit would gain in value over time if they kept banning automation, because it is increasingly difficult to avoid AI-generated material polluting your dataset, no matter how much you avoid automation and try banning it. Inevitably, some AI-generated material is going to get in.

      It’s a problem in two ways:

      1. The vast vast majority of data on Reddit has already been sold, so you can’t rely on that data for future revenue
      2. The remaining data that’s current is polluted by AI and is therefore worth less than the historical data because the more AI pollutes your dataset, the more likely it is to lead to Model Collapse, where an LLM is poisoned due to unverified data generated by other LLMs

      I am firmly of the belief that sites like Internet Archive will be some of the most valuable companies in the AI space, because they hold an immense amount of untainted data created prior to 2019.

      • @[email protected]English
        1•19 days ago

        Except AI companies are already using data from Internet Archive wholesale with paying them a cent.