Arpy Dragffy Guerrero
speaker
627 appearances
7 recordings
1 series
first heard May 2026
last heard 6 Aug
Arpy Dragffy Guerrero’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 7 in all, peaking in Jun 2026 with 3.
Appearances
Now, before we get to consistent adoption and real value creation, there are a few things that have to get solved that haven't really been yet.
Now, the 2026 struggle is real.
The era of encouraging everyone to build everything with AI and just token max, use as many tokens as possible, is coming to a close.
It is simply too expensive.
And it doesn't show the same returns as an initiative for all workers that brings the flora.
More than that, we're coming to terms with the reality that most vibe-coded applications, these are quickly assembled AI-assisted builds without proper engineering review, are like a house of cards.
They open businesses up to serious security risks and governance and security, so deciding who can use what, how outputs get reviewed, what data gets shared, are essential.
So the data is already showing that leadership who are overly gung-ho with AI are at least twice as likely to security risks.
This is not acceptable.
We have to figure this out.
So unfortunately,
Most businesses have not properly thought about governance and security or are looking at governance as being a blocker as opposed to what it actually can be, which is guardrails and framework to allow people to run faster.
These teams are not equipped to manage an entirely new work stream like that.
Most of all, very few businesses have something like a chief AI officer.
Without someone in that seat who has the accountability and the mandate and the budget to be able to make sure that an organization's AI strategy is coherent and is working towards true business value, you just end up with a collection of independent experiments with no real connective tissue and a legacy of
unusable agents and tools and of course there's always the risk of being replaced by ai right that's what i'm hearing from people at meta and many others who have been worried about as their work and actions are being tracked to train these systems what will it mean for them
But the more real situation happening inside organizations is this.
The more we build with AI, the more we realize how unreliable most AI systems actually can be without a human watching over them.
So Dan Shipper, the founder of Every, and honestly, one of the most clear-eyed writers on how AI actually works in practice, captured this on a recent episode of Lenny's podcast.
And honestly, this is a conversation that Arpi and I have quite often too.
Showing 481–500 of 627 · page 25 of 32
← Previous
Next →