New FLiRT COVID variants emerge + 1 more story

Thank you for everyone's responses in the last poll. See the results after the news.

Today ChatGPT read 1081 top news stories. After removing previously covered events, there are 2 articles with a significance score over 6.

[6.6] Microsoft and OpenAI launch $2M fund against election deepfakes — TechCrunch

Microsoft and OpenAI have launched a $2 million "societal resilience fund" to combat AI misuse, such as deepfakes, which threaten to deceive voters in the year's global elections involving 2 billion people across 50 countries. This initiative, part of a broader commitment to responsible AI, includes funding organizations aimed at raising AI literacy among voters, especially in vulnerable communities. They have also introduced technology to identify fake content and have joined an industry group focused on content authenticity.

[6.3] New COVID-19 variants FLiRT spreading globally — The Hindu

New COVID-19 variants, KP.2 and KP1.1, are causing renewed concern worldwide with cases rising in the U.S., U.K., South Korea, New Zealand, and India. Identified as FLiRT variants, they show greater resistance to current vaccines and prior infections.

Highly covered news with significance over 5.5:

[5.6] AstraZeneca withdrawing Covid vaccine due to rare side effects
(The Telegraph + 4)

[5.5] Ukraine arrests two security officials over plot to assassinate Zelensky
(The Telegraph + 6)

[5.5] The alleged LockBit ransomware mastermind has been identified
(WIRED + 4)

Wow, you do not make it easy! The votes split almost 50/50.

The question was, “Should we decrease the significance rating of the "highly covered news" section from 5.5+ to 5.0+?”

With this split, it’s important to look into the reasons behind the votes. I read all the comments, and here’s how I see the groups:

• The “keep as is” group seems to value minimalism more. They don’t mind sometimes missing something if that means that most of the time, all the remaining stories are significant and relevant.

• The “more news” group seems to value coverage more. They don’t mind seeing some irrelevant headlines because they could easily skip them.

I understand both groups. And I think there’s a way to improve things for everyone.

I simply need to improve the scoring. Many people in the “more news” group said that some of the articles between 5.0 and 5.5 were “important”, “interesting”, “newsworthy” and “relevant” and were given scores that are too low. If they were cut off, it means that the algorithm is not ideal and can be improved further.

So that’s what I’ll work on. I already have a couple ideas: new certainty parameter and rework of the potential parameter.

And there’s another we could do.

One user left a simple and genius comment:

— “Does it have to be a 0.5 decrease?”

7 out of 8 additional stories on that day were actually in the range from 5.0 to 5.2 and wouldn’t be included if we had a cutoff at 5.3+. With that threshold, we have only two additional stories today:

[5.3] TikTok sues US government over potential ban, citing First Amendment (The New York Times + 4)

[5.4] Japan captures first image of space debris from orbit
(Livescience.com + 4)

They seem good enough to me. So let’s try it?

One last thing: I’ve enabled comments under the posts in web version. Today seems like a good day to try them. Scroll to the top of the email and click “Read online”. Comments section will be at the bottom (you might need to log in).

Thank you again and until next time!

Vadim