4.7 • 984 Ratings
🗓️ 9 April 2025
⏱️ 20 minutes
🧾️ Download transcript
Google unveils its new TPU’s, this time with inference! Temu and Shein officially get nuked from orbit. More signs of pullback in AI datacenter buildout. Are we actually, for real, about to get an iPad Instagram app? And a cute little home robot from ages ago, looks like it’s finally coming to a house near you this summer.
Sponsors:
Links:
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.
Click on a timestamp to play from that location
0:00.0 | Welcome to the Tech Meme Right Home for Wednesday, April 9th, 2025. I'm Brian McCalla today. Google unveils its new TPUs, this time with inference. Tammu and Shian officially get nuked from orbit. |
0:14.3 | More signs of pullback in AI data center buildout. Are we actually for real about to get an iPad Instagram app? And a cute little |
0:22.0 | home robot from ages ago looks like it's finally coming to a house near you this summer. |
0:26.4 | Here's what you miss today in the world of tech. |
0:33.0 | Google's Cloud Next conference is going on this week and there. Google unveiled Ironwood |
0:38.5 | its seventh generation TPU and the first it's ever made for inference, coming later in 2025 in a |
0:44.6 | 256 chip cluster or a 9,216 chip cluster, quoting TechCrunch. Ironwood arrives as competition |
0:53.1 | in the AI accelerator space heats up. |
0:55.7 | Invidia may have the lead, but tech giants, including Amazon and Microsoft, are pushing their own in-house solutions. |
1:01.0 | Amazon has its tranium, inferentia, and graviton processors available through AWS, and Microsoft hosts Azure instances for its Cobalt 100 AI chip. |
1:11.6 | Ironwood can deliver 4,614 T-flops of computing power at peak, |
1:17.1 | according to Google's internal benchmarking. |
1:19.4 | Each chip has 192 gigabytes of dedicated RAM with bandwidth approaching 7.4 T-BPS. |
1:26.2 | Ironwood has an enhanced, specialized core, sparse core, for processing the types of |
1:30.9 | data common in advanced ranking and recommendation workloads, e.g. an algorithm that suggests |
1:35.8 | apparel you might like. The TPU's architecture was designed to minimize data movement and |
1:41.1 | latency on chip resulting in power savings, Google says. |
1:44.7 | Google plans to integrate Ironwood with its AI hypercomputer, a modular computing cluster in Google Cloud |
1:50.1 | in the near future. They added, end quote. |
1:53.5 | Google also launched Workspace Flows, an AI tool to automate multi-step processes in Alpha, |
2:00.1 | and new docs, sheets, meet, chat, and vids |
2:03.3 | features like audio in docs, quoting 9 to 5 Google. As first teased at I.O. last year, the company |
... |
Transcript will be available on the free plan in 4 days. Upgrade to see the full transcript now.
Disclaimer: The podcast and artwork embedded on this page are from Ride Home Media, and are the property of its owner and not affiliated with or endorsed by Tapesearch.
Generated transcripts are the property of Ride Home Media and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.
Copyright © Tapesearch 2025.