Safety & governance fell 13.9 points and Infrastructure rose 6.1 points.
Where it showed up
Safety & governance and Infrastructure accounted for 11% of the relevant conversation across 21 episodes.
How to read it
Recurrence across podcasts does not mean agreement. The cited moments show the language behind each summary.
In the discussion
First of all, you have to trust what the AI is telling you. And that's where explainability is important. And our platform's always been explainable from the very beginning. It's not a black box.
Ways to reduce harm, set limits, test behavior, and govern how AI is used.
Trend this week
Losing attention3%of conversation
2 of 2 published weeks met the cross-podcast threshold.
11quoted moments6episodes3podcasts discussing this topic
A closer read
What changed
Its share fell from 16.5% to 2.6%, with evidence across 6 podcasts.
Where it showed up
It appeared in episodes including “Big Pharma Fails 50% of the Time in Phase Three. AI Can Fix That | Vin Singh, BullFrog AI” and “The Big Ways AI Just Changed”.
How to read it
6 cited episodes from 3 podcasts are shown below. The topic appeared across 6 podcasts in the full analysis.
Podcast moments from the week
11 quoted moments · 6 episodes · 3 podcasts
Eye on AI
Big Pharma Fails 50% of the Time in Phase Three. AI Can Fix That | Vin Singh, BullFrog AI
“First of all, you have to trust what the AI is telling you. And that's where explainability is important. And our platform's always been explainable from the very beginning. It's not a black box.”
“This is a project that I had had kicking around for weeks at that point. And because Fable 5 came around, I just decided to get it done and done it got in one fell swoop. And thank goodness, because as we know now, Fable 5 wouldn't be around all that long. And in those first c...”
“Michael claimed the software will tell Anthropic whether a user is in China or even if a user is associated with a Chinese lab. This information was extrapolated from timezone settings and other metadata. Now, we knew Anthropic was doing something to monitor the activities of ...”
“On these foundations of trustworthy AI, we talked about live data, the right data, and then guardrails. Can you talk about the guardrails that are necessary, why they're necessary?”
“ensuring that the prompts are safe, protecting yourself against prompt injection, forbidding the agents the access to, I don't know, to certain tools”
“The challenge of enforcing these guardrails and governance across distributed environments, how do you handle that? And how does agentic AI make that even harder?”
“My gratitude to companies across industries who continue to work closely with the White House to implement the president's executive order promoting advanced AI innovation and security. This includes excellent work around advanced model access and guardrail testing and securit...”
“Including Claude, Haiku 4.5, Sonnet 4.6, Opus 4.6, Opus 4.7, Opus 4.8, GPT-5.4, GPT-5.5, and Kimi K 2.7. Importantly, they continued, the reported technique did not expose any unique Mythos-level cyber capabilities. The behavior reflected a borderline case for Fable 5's safegu...”
Its share fell from 16.5% last week to 2.6% this week.
Share of this week’s podcast conversation
Recent weeks16.5%→This week2.6%
Came up less often than in recent weeks.
Podcast moments5 cited momentsViewClose
Eye on AI
Big Pharma Fails 50% of the Time in Phase Three. AI Can Fix That | Vin Singh, BullFrog AI
First of all, you have to trust what the AI is telling you. And that's where explainability is important. And our platform's always been explainable from the very beginning. It's not a black box.
This is a project that I had had kicking around for weeks at that point. And because Fable 5 came around, I just decided to get it done and done it got in one fell swoop. And thank goodness, because as we know now, Fable 5 wouldn't be around all that long. And in those first c...
Michael claimed the software will tell Anthropic whether a user is in China or even if a user is associated with a Chinese lab. This information was extrapolated from timezone settings and other metadata. Now, we knew Anthropic was doing something to monitor the activities of ...
AI Agents Are Failing and It's Almost Never the Model's Fault | Alberto Pan, Denodo
On these foundations of trustworthy AI, we talked about live data, the right data, and then guardrails. Can you talk about the guardrails that are necessary, why they're necessary?
Okay. So one challenge obviously with this other trillion dollars worth of compute that's going out into the world is it's super heterogeneous, right? There's many, many different devices, different chips, et cetera, et cetera. So how much of the work that you're doing is abou...
But Claude Tag wasn't just another way to interact with Claude via Slack, It was instead a way for people in any part of Slack to call upon the power of Claude Code. This democratizes access to the advanced technical capabilities of Claude Code. It gives Claude Code access to ...
On Truth Social, Trump declared Micron a, quote, truly great American company, one of the hottest anywhere in the world, adding that this is, quote, the biggest corporate investment of its kind and will help jumpstart the American dream for these fabulous children as we celebr...
Sources said the plans are still in development and subject to change, but they're far enough along to have a name and a leadership team in place. The business line will be called Meta Compute and is led by Head of Infrastructure Santosh Jennerhan working alongside Daniel Gros...
And yet, especially in larger organizations, the orchestration layer of figuring out how all these new disparate pieces work together, especially as the work output increases dramatically, is going to be extremely important. Orchestrators that live at the intersection, not jus...
Intelligence on the Edge: Liquid AI's Ramin Hasani on the Search for Device-Native Foundation Models
Right? So in those kind of places, you would go towards extremely biased and even maybe just differential equation-based models like liquid neural networks today. They're getting applied to many, many different applications. I see that because the repository, the original repo...
And so far I've been okay Okay with that. The benefits are certainly worth whatever risk I'm taking, I feel. But I would maybe love to run that data through a local model first so that I don't have to send all my data to the cloud every time. Now, notably, what it gets filtere...
Fable did only 20% of the work. WTF, disappointing. Is anyone else seeing this? On the flip side, Bridgemind also noted that quite a lot of their work was being routed to Opus, commenting, I just paid $321 for a coding session where Fable 5 refused to do the work. And yet othe...
Its share grew from 2.4% last week to 6.2% this week.
Share of this week’s podcast conversation
Recent weeks2.4%→This week6.2%
Came up more often than in recent weeks.
Podcast moments5 cited momentsViewClose
The Cognitive Revolution
Intelligence on the Edge: Liquid AI's Ramin Hasani on the Search for Device-Native Foundation Models
Yeah, but let's zoom in maybe a little bit more on that neuron. And I guess I'm wondering a little bit too, what is the fundamental limit, or what are the first bottlenecks that you you hit when you try to scale the original liquid approach. With the sigmoid-type functions tha...
Just for calibration, when you talk about the liquid neural networks that have gone up to hundreds of thousands or pushing a million neurons. What kind of hardware does that run on? Is that like a CPU supercomputer?
The Journal described it as a, quote, handset-like device designed to reshape how humans interact with AI. They wrote that it's slimmer than an iPhone, would run on a proprietary operating system, and would integrate technology from xAI. Elon was quick to deny the reporting, c...
Image Generation and Visual Intelligence with Black Forest Labs
That's awesome. And just to follow up on that, I know some people out there, our listeners are always trying things on, you know, on their laptop or wherever they're, they're, they're pulling down things. I know for quite some time when I tried to access some of these models a...
Its share grew from 18.5% last week to 21.6% this week.
Share of this week’s podcast conversation
Recent weeks18.5%→This week21.6%
Came up more often than in recent weeks.
Podcast moments5 cited momentsViewClose
Eye on AI
Big Pharma Fails 50% of the Time in Phase Three. AI Can Fix That | Vin Singh, BullFrog AI
Yeah. The, at the last stage there, when you said then you have these different targets competing with each other, are you using evolutionary AI where you're You're coming, different models produce an output and then you take the best and combine those models. Yeah. Yeah.
Its share fell from 30.1% last week to 27.8% this week.
Share of this week’s podcast conversation
Recent weeks30.1%→This week27.8%
Came up less often than in recent weeks.
Podcast moments5 cited momentsViewClose
Eye on AI
Big Pharma Fails 50% of the Time in Phase Three. AI Can Fix That | Vin Singh, BullFrog AI
Yeah. And also, you're talking about big pharma, but what's happening are there are smaller companies, labs that are developing molecules or drugs, and they bring them all the way up through phase 1, phase 2 trial, and then sell them to big pharma. Is that affecting that strat...
Intelligence on the Edge: Liquid AI's Ramin Hasani on the Search for Device-Native Foundation Models
We are ranked number 5, actually, in the US from the number of downloads. the popularity perspective, we have over 1 million downloads per week on Hugging Face. That's a lot of downloads of these small models that we are building. They're very popular. The top ones from the US...
Okay, so let's fast forward to the present. You guys are now in the market working with customers. One notable conversation I heard with one of your customers was with the CTO of Shopify, by on the Latent Space podcast, who had some very good things to say about you and your t...
Uber made headlines when it set a $1,500 per month cap on AI spend. And these stories and the others like them reinforced the idea that token efficiency and token discipline were going to become important new aspects of the AI landscape, particularly in the enterprise. And sta...
