S3 E3: Byte: The internet of lots of things
No, Facebook isn't listening to you shout about vegan pizza, and I spent 48 hours proving it.
- Smartphone improvements are now marginal like cars, with software lagging behind hardware because developers can't fully optimise for one OS before the next arrives
- 5G, Wi-Fi 6 and edge computing tackle latency and power draw, which is what actually unlocks persistent connected devices and use cases like driverless cars
- Chips have to carry 2G through 5G simultaneously during the transition, which hurts battery life and dampens early impressions of new networks
- The next wave of data will be machine-generated rather than human-triggered, giving businesses far more granular event-level data for analysis
- Targeted ads come from behavioural patterns and shared signals like a common router MAC address, not from your phone secretly listening to you
This episode argues that we're at 'peak smartphone' in hardware terms, and that the real next wave of innovation and data comes from connected devices working together via better networks (5G, Wi-Fi 6) and edge computing — not from your phone secretly listening to you.
A conversational podcast between two hosts covering smart home devices, wearables and network infrastructure, framed around 'the internet of lots of things'. No hands-on demo; it's a discussion of trends and a debunking of a common privacy myth.
- Smartphones have hit a hardware ceiling 2:51
Camera, battery and screen improvements are now marginal, much like cars. The bigger problem is that software never catches up to hardware because a new device generation arrives before developers can fully optimise for the last one.
- Smart home devices only work if they connect 6:52
Every smart speaker, thermostat or plug depends on some wireless protocol to talk to something else, and that dependency on Wi-Fi setup and network reliability is often the weakest link in the experience, not the device itself.
- Latency, not speed, is the real bottleneck 10:13
Wi-Fi, Bluetooth and mobile networks all suffer from latency, which is why crowded places like stadiums degrade connections. This is the specific problem 5G and Wi-Fi 6 are built to solve, enabling persistent low-latency connections rather than just faster downloads.
- New chips have to support every prior network generation 15:02
A 5G-capable phone still has to carry 2G through 4G support because coverage isn't universal yet, and running all of that simultaneously is what drains battery life and makes early 5G phones feel disappointing — expect a few years of this trade-off before it settles.
- Edge computing shifts power and processing off-device 18:28
Offloading computation to the cloud (as with cloud gaming) or using specialised on-device chips (as with tablets outperforming laptops on specific tasks) reduces what a device needs to do locally, which is where real efficiency and capability gains are coming from now, not raw chip speed.
- Wearables and smart devices should augment, not just automate 23:41
The most useful framing for a connected device isn't 'it does the lazy thing for me' but 'it removes friction so I can spend that time on something better' — worth applying when evaluating whether a smart feature is actually adding value or just novelty.
- Machine-generated data is about to dwarf human-triggered data 28:22
Historically most business data came from a human action (a click, a purchase). Connected devices generate continuous event-level data independent of a human trigger, which means far more granular data for analysis than businesses have had before — a shift worth preparing for if you work with data.
- Debunking the 'my phone is listening to me' myth 30:02
Targeted ads come from aggregated behavioural signals, not audio surveillance — image recognition on your phone happens locally and doesn't identify you personally, and coincidences like shared adverts across friends are usually explained by a shared Wi-Fi router's MAC address, not eavesdropping. The real privacy concern is how securely your data is stored, not whether someone is listening.
- Devices on the same Wi-Fi router share a MAC address, which is a plausible explanation for 'coincidental' shared ads between people who live or work together — not evidence of listening.
- Image recognition in phone photo apps processes pixels as numbers locally; it isn't a human or a company looking at your face.
- 5G rollout means phones must simultaneously support older network generations, which is why battery life often gets worse before it gets better.
- The bigger real privacy risk raised is data security and breaches, not passive listening — an argument the hosts make is under-discussed.
Useful framing when you're evaluating a new connected device or wearable purchase, explaining to someone why targeted ads aren't proof of eavesdropping, or thinking about why more granular machine-generated data is coming to your analytics work.
How this Rollup was made provenance & method
A Rollup is drafted by AI from the video's transcript, then reviewed and edited by Tim. Everything used to produce this one is listed below (the model, the exact prompt, and the source video), so the process is transparent and reproducible.
- Transcription
- On-device: NVIDIA Parakeet v3 for recent videos, OpenAI Whisper large-v3 for earlier ones. The transcript never leaves the machine or gets published.
- Drafting
- Claude Sonnet 5 in the cloud, from that transcript.
- Prompt
- The exact Rollup prompt (v2) , the full system prompt, unedited.
- Source video
- Watch on YouTube
- Drafted
- 5 July 2026 at 09:38
- Reviewed & edited
- 10 October 2026 at 14:41 · by Tim Ngwena
Model + prompt + video is everything you'd need to recreate a Rollup like this yourself. The one thing we don't share is the transcript.
Rights. The video and its transcript are the property of TN Media Ltd. Unauthorised use or download is prohibited. © TN Media Ltd.
