Jul 20, 2026 · 31m · big-technology
Inside The Rise of Physical AI — With Amir Khoshniyati
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Alex Kantrowitz interviews Wiliot VP Amir Khoshniyati to explore how battery-free ambient IoT sensors and Physical AI transform supply chain logistics, eliminate perishable waste, and deliver measurable enterprise ROI.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 30.7% of the talking time here. How this is scored →
Alex as informed peer 4.3 Guest teaching 3.8 Guest disagreement 0.6 Alex pushing back 2.3
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 20:13 Amir defends automated infrastructure against failed vision rollouts
Amir firmly pushes back on the relevance of Starbucks' failure, arguing that friction-heavy manual vision systems fail while frictionless ambient tracking succeeds.
Hardest push from Alex ▶ 19:29 Alex challenges autonomous supply chain claims using StarbucksAlex directly rejects Amir's optimistic timeline on autonomous agents by bringing up Starbucks' high-profile inventory AI breakdown.
Biggest teaching moment ▶ 26:30 Amir breaks down the $270M retail delay calculationAmir educates the host with concrete enterprise metrics, demonstrating how a 37-minute pallet dwell threshold translated into a $270 million annual loss.
Alex holds their own ▶ 19:29 Alex injects real-world deployment failure evidenceAlex demonstrates subject expertise by countering theoretical autonomous agent timelines with a specific, reported enterprise failure in retail computer vision.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Tracing the Transition from Connected Devices to Physical AI | 4 | 4 | 1 | 1 | Alex opens by detailing Amazon's Hands Off the Wheel algorithm from his book Always Day One. Amir traces the historical shift from internet-connected devices to ambient physical AI giving physical items digital identities. | |
| Real-Time Environmental Sensing Beyond Traditional Spreadsheets | 4 | 4 | 0 | 1 | Alex frames the shift away from manual inventory spreadsheets to real-time telemetry. Amir explains Williot's postage-stamp-sized, battery-free IoT pixels that harvest ambient RF energy. | |
| Transforming Raw Sensor Telemetry into Intelligent Predictions | 4 | 4 | 1 | 3 | Alex pushes Amir to clarify where artificial intelligence actually comes into play beyond basic telemetry data collection. Amir breaks down predictive dwell times and conversational item queries regarding food safety. | |
| The True Return on Investment in Industrial and Physical AI | 5 | 3 | 0 | 1 | Alex articulates the high ROI of industrial AI compared to consumer subscription models and asks whether AI can make autonomous supply-chain decisions. Amir illustrates this with proactive temperature and bacterial tracking in produce. | |
| Overcoming Probabilistic AI Hallucinations with Ground Truth Data | 5 | 3 | 1 | 4 | Alex questions how enterprises can rely on probabilistic, hallucination-prone AI models in mission-critical food safety contexts. Amir explains that grounding AI in verified physical sensor feeds eliminates hallucinations. | |
| Evaluating Autonomous Supply Chains and Implementation Pitfalls | 6 | 4 | 2 | 6 | When Amir suggests autonomous agents could soon run supply chains, Alex mounts a strong pushback by citing Starbucks' failed computer vision shelf-inventory deployment. Amir responds by contrasting manual friction with frictionless ambient sensing. | |
| Multimodal Foundation Models and Cross-Supply-Chain Intelligence | 3 | 4 | 0 | 2 | Alex explores multimodal AI trained on physical telemetry, and Amir outlines how shared upstream data pools across common distributors enable cross-model predictive learning. | |
| Measuring Hard ROI: A $270 Million Dollar Retail Case Study | 4 | 5 | 0 | 2 | Alex asks how organizations quantify AI return on investment amidst token spend concerns. Amir shares a case study of a major retailer uncovering $270 million in annual losses caused by 37-minute pallet dwell delays. | |
| Integrating Agentic AI Models with Enterprise Systems | 4 | 3 | 0 | 1 | Alex asks about token consumption trade-offs in agentic versus analytical workflows, and synthesizes the physical data foundry metaphor as raw inputs transforming into refined outputs. |
Statements from this episode (11)
Kantrowitz: Amazon Replaced Vendor Managers with AI 'Hands Off the Wheel' System
“And they have this program called Hands Off the Wheel, where they basically took past purchase data, things like what people bought, and what zip codes, and what season, and predicted how to stock their fulfillment centers with what products, when, and at what…”
Khoshniyati: Early IoT era was an 'Internet of Devices', not things
“Internet of Things, when it launched, and it was before the COVID times, it excited everybody, but truly the understanding of it really wasn't stabilized yet. So everyone was using the coin term, But the reality was we were just moving from every type of devic…”
Khoshniyati: Physical AI gives every asset digital identity and condition tracking
“Physical AI is our ability to take every physical asset in the world and give it a digital identity. And you go beyond just having recognition of what information from that asset is associated to a tag around it, but the condition that that asset is sitting in…”
Khoshniyati: Physical AI Value Lies in Condition Telemetry, Not Just Location
“And it's beyond just knowing the location and the traceability of that item, but the condition that it's in. And that's where physical AI starts to pick up value beyond what historically was in place. So with condition, you know the temperature of that item. Y…”
Khoshniyati: Wiliot IoT Pixels Harvest Ambient Waves to Transmit Telemetry Data
“We have proprietary IP around IOT pixels. So these are, ah, reference designs that all RFID manufacturers could produce. These tags go on physical items. They're battery free, so they harvest energy off of ambient waves. So we have a capacitor that acts like a…”
Khoshniyati: Physical AI eliminates finger-pointing between food processors and receivers
“It creates a frictionless handshake between food processors and the folks that are receiving the items. So there's no finger pointing. You can actually go to the source of the truth where something went wrong and then fix the compliance between those two parti…”
Khoshniyati: Wiliot sensors detect spoilage and halt damaged produce mid-transit
“We've had these examples with everyday items like strawberries that go through some level of, you know, refrigeration. They come out, there's condensation. It goes through another temperature variance that it might be too cold, and then there's freezing. And w…”
Khoshniyati Predicts AI Agents Will Soon Run Global Supply Chains
“With the right infrastructure behind it and process I don't think we're far from it.”
Khoshniyati: AI Models Are Useless Without Ground-Truth Physical Tracking Data
“If you don't have a reliable data source, And the assets slash products that you're tagging, they're not providing the right data triggers behind it and the right insights into those assets and products when they're through the supply chain. None of this matte…”
Khoshniyati: Major Retailer Lost $270M Annually Due to 37-Minute Pallet Delays
“We have a major retailer out there. They factored in that every time they had a pallet and it was sitting idle more than 37 minutes, there was loss. And then we multiplied that out on an annual basis. And that 37 minutes translated to north of two hundred seve…”
Khoshniyati: Standardized Physical AI Telemetry Will Provide Baselines for Small Businesses
“But as we get macro level and this becomes standardized, you will start to pick up the trends, very similar to the example earlier, that food processors that work with distributors are going to have commonalities. And then some of these more public domains aro…”