Is AI Moving Too Fast to Control? artwork

Is AI Moving Too Fast to Control?

The Daily AI Show

July 29, 2026

The episode focused on new details from the OpenAI and Hugging Face security incident, including additional services accessed by the models, an Artifactory zero-day vulnerability, and the ability of AI agents to find exposed credentials from older breaches.
Speakers: Beth, Brian, Andy
**Beth** (0:00)
Hey, what's going on, everybody?

**Brian** (0:01)
Welcome to The Daily AI Show. Today is July 29th, 2026
It's episode 778 And with me today are Beth and Candy, and I'm Brian, and we are live. We're live every weekday. If you're just finding us, which that happens all the time, we still have lots of people who aren't subscribed and coming back. So if you're new to us, just know that we do this Monday through Friday. As you heard me say, 778 episodes as of today. We're coming up on our three-year anniversary in about two Fridays from now, a week from Friday.
Then we'll hit our 800th anniversary in August. So lots going on here. We've been doing this for quite a while, having the day-to-day conversations about AI, talking about all the latest news, but also just talking about what it means.
Where are we in this great AI journey? What does that mean for you as the listener? Ultimately, it's we want you to know that you're not behind. We started the show so that we could all collectively have a conversation together. One of the great ways that you can do that is there's a couple of ways, but if you want to literally join us live on the show, then the best way to do that is inside the YouTube. Go to our YouTube channel, excuse me, do it for the live episode and you'll be able to join the comments, which people do all the time. We actually have a great set of people who are in there most days. Very friendly, very inviting, so you're welcome there. If you can't make it live at 10 AM Eastern, totally understand that. Then we also have a community and you can go to the dailyaishowcommunity.com in order to sign up for our free Slack community. We have lots of great conversations in there as well. All sorts of ways that you could be part of the show, as well as our newsletter that goes out on Sundays. That's another great way to get some of the stories. The way the newsletter basically works is we'll pick some of the major stories, the major themes throughout a week coming from the shows, and then we'll rework them and look at them from different angles. Because obviously, we do live, me, Beth, Andy, and we'll all do live conversations. We'll do quick fact checking as we can on the show. But oftentimes, these are great conversations. We don't have necessarily time to go deep into them. And so the newsletter usually will pick three of those main topics that myself or Andy feel like are really interesting. And then we'll put those out to everybody in the newsletter. So there you go. There's all the ways you can interact with the show. Obviously, a like or subscribe is always super welcome. But we want you to feel welcome here first. I feel like this is the place for you. So if you do feel that way, hey, a like or subscribe really always does help. But we want to earn it from you. Okay, guys, I know one of the updates from yesterday that I don't... I mean, we talked about this. We were actually talking about how some of these news stories seem to be staying in the news feeds. This one, though, actually has a good reason for it. So Andy, you were even saying, we were talking about before anybody else turned yesterday, sort of like the troughs and the crest of the waves. And we sort of feel like we're in this trough of like, there's not like a ton of information. I've been digging a little bit deeper. I've been pushing on perplexity. In fact, I was just arguing with perplexity before we went live saying like, no, I feel like there's more to uncover here. But one of the things that is worth definitely bringing up is that the OpenAI and Hugging Face security incident is not over. In fact, we're getting these little trickles out of information about more things that happened during the security breach from OpenAI, I should say.
And what they did is yesterday, they put an update on top of the blog that they had. So this is from openai.com. It's called OpenAI and Hugging Face Partner to Address Security Incident During Model Evaluation. That's what the article is called. But they added a blurb to the top that says updated on July 28th, that would be yesterday.
So I you know, look, there's there's about seven or eight points here. One of the things that stood out to me and you know, Beth, I saw you shaking your head as well. So you might have more to do this is that there are more sites that were accessed than previously understood.

46 more minutes of transcript below

Feed this to your agent

Try it now — copy, paste, done:

curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000778979182