The AI Round-up: Hopes > Nopes

The AI Round-up: Year 4, Issue 33

We’re starting to see some steps being taken to protect important things in the age of AI…like jobs, journalism and our mental state.

Today’s issue is going to take a look at those things (and more) as we pivot to a ‘lighter’ round-up for a week.

Why? Well, because if you want to stay plugged in you are seeing a mix of about 80/20 (80% of stories skewing doom), and I think we could all use a break.

Let’s get to it.

Non-profit to spend $10M on journalism.

Microsoft announces its plan to build SI for good.

Two things: first, that’s my interpretation of Satya Nadella’s annual letter. Second, HE is calling it super intelligence, not me. Hence, the SI (v. AI).

Sam Altman outlines three AI scenarios. Includes a good one, he seems to like, and a bad one, he calls terrible.

We may be entering the age of compliance sooner than later. All things considered…that is a good thing.

Note: the linked article is a summary of the reports/interviews done after this Altman meeting, first reported on by The Information. The Information doesn’t give freebies and is an expensive subscription, so I linked to the Medium summary article instead.

Muse is caught sharing personal details…and Meta actually paused that function. (The Verge)

See, I told you today would be all positive stories.

Christopher Penn releases Highly Opinionated AI Skills Fixer.

Christopher Penn knows how to name things.

Speaking of Christopher Penn, he broke my brain a bit with his thoughts on how we should talk to our LLMs. And guess what, it’s using that language we have grown to hate/spot.

Anthropic proves…you need to check your work.

So this study is pretty fascinating in the sense that it shows how analysts use/weigh insights from their AI agents. They are more likely to lean the way their agent is suggesting, especially early in the process. But not as much in the final decision. Meaning? We, the human, are still in the loop.

Google loses $500B in value. (BBC)

Wait, I said only good things this edition.

Right. Which is why this story is here. This, is a good thing…if you believe Alphabet/Google CEO Sundar Pichai.

DoorDash uses AI to cut reorder times from 16 seconds to 2 seconds. (InfoQ)

Finally! Now, what are you going to do with that extra 14 seconds? Whatever you feel like, gosh!

The first fully AI-generated movie flops. (The Hollywood Reporter)

Ha-ha!

 

Some Nopes.

 

1. Bank of America is warning “don’t give your banking information to AI agents”. (Shelly Palmer). No, seriously – don’t.
2. AI layoffs are starting to look familiar. (Fortune) If you want to believe 40 years of research.
3. A code red in the hacking/cyber security space. (ABC News)
4. The Pope warns human decisions should never be delegated. (The Vatican) Unless they’re fantasy football projections.

Live life, love life,

Ben

As a reminder, this is a round-up of the biggest stories, often hitting multiple newsletters I receive/review. The sources are many … which I’m happy to read on your behalf. Let me know if there’s one you’d like me to track or have questions about a topic you’re not seeing here.