Matt Campbell’s Post

The value of posting on the internet, in one image

Excited to announce a newsletter that Amean Asad and I are starting, where we explore the ways we consume content and how we use that knowledge to apply to our daily lives. The Commonplace will be the medium where we distill our thoughts, research, and ideas around knowledge discovery and retention, in the age of information. Check out our first article discussing the shift of content from physical to digital, and the problems that come with it. Very excited to hear your thoughts as we explore this topic space! https://lnkd.in/eGdvCicD

Exciting new insights on social biases in text-to-image foundation models! Our latest blog post delves into the impact of these biases on downstream applications such as image editing and classification. We uncover significant inter-sectional biases in a leading text-to-image model, urging caution in adopting such models for practical tasks. Read the full study at https://bit.ly/4at9D2M.

📢 [Publication Alert] I'm thrilled to share our latest research, "Cracking Open the European Newsfeed," co-authored with Luca Rossi and Fabio Giglietto, now available on JQD Digital Media https://lnkd.in/d5WRfere This study marks a pioneering effort in longitudinally measuring the circulation and exposure of low-quality information on Facebook. Our focus was to track the prevalence of low-quality news domains on Facebook in Italy, France, and Germany from 2019 to 2022. This investigation led us to several key findings: 1️⃣ General Decline in News Domains. We observed a significant decrease in the number of news domains shared on Facebook across all three countries. Interestingly, this trend was evident even though France and Italy have higher Facebook adoption rates compared to Germany. 2️⃣ Decline of Untrustworthy Sources. The proportion of low-quality news sources out of the total news sources has decreased in France. However, it has risen in Italy and Germany, notably during election periods. 3️⃣ Views of Untrustworthy Sources. Despite the increase in low-quality news sources in some countries, the overall views of untrustworthy sources have diminished. This is indicative of effective measures in reducing the reach and visibility of such content, with Italy in 2022 being a notable exception. These insights are crucial in understanding the dynamics of problematic information dissemination on social media platforms.

Guangyi Li has recreated a web schematic of what your web experience is like today, which includes the worst of the worst, from out-of-place ads to abusive pop-ups, triggering notifications, chats, ratings, content prohibited by regions… The coolest thing, come on. It is a great example of the shittification (as coined by Cory Doctorow) that the Web and (almost) all the sites that inhabit it have reached, sometimes due to inevitable drifts, sometimes due to "legal issues" or pressures of some kind. It is something that not only occurs on the web, but also in apps on platforms such as Netflix, Twitter, Instagram or Amazon. We have become accustomed to living with it, assimilating it into the culture of the companies and among the people who roam the Internet, "unpresentable and hateful products, the result of going through the roller and the dictatorship of analysts, markets and excessive ambition”. Follow the link below and you will find the experience terribly familiar: https://lnkd.in/eu4ABEWG

In this blog post, I'll delve into the intriguing debate between Closest Friends Instagram Only and the Public Feed. Read more 👉 https://lttr.ai/AOqRQ

In this blog post, I'll delve into the intriguing debate between Closest Friends Instagram Only and the Public Feed. Read more 👉 https://lttr.ai/AM3RL

Why bots are useless number 1,001 If I want to bulk upload a list of invitees to an event in #Google Calendar, I have to create a group. When I create a group. there is an unspecified maximum number of new members (both the number and the fact of the limit is unspecified until after it fails) and when it fails, it fails in its entirety. But it doesn't specify that, either. There's a complaints page. I added mine. This is it: "Me, too, but it didn't even tell me what the limit is or how it can be changed. It seems that if the copy/pasted list is longer than the number allowed Google sorts it alphabetically, then rejects the entire list. " The bot responded "failed to post. See our community guidelines." "#communityguidelines" are no help. So, that's it. It's just too much trouble which means that for news of The Financial Crime Forum, you'll just have to subscribe to the mailing list at https://lnkd.in/g_vnJCvJ which is where invitations and confirmation of bookings will be sent from. That wasted a whole morning that I had thought I was investing in future time-saving. Grrrr. Tell me again how #BotTech is going to replace humans... #morriscotterill #IT #ai #bots #actualintelligence.