ImageNet, every mention
16 scenes, the whole family · ← back to ImageNet
tap a year for its mentions
every year anyone Matt Turck 3Richard Socher 2Benedict Evans 2Antoine Bordes 2Yann LeCun 1William Falcon 1Serkan Piantino 1Sanjit Biswas 1Neil Zeghidour 1Mike Abbott 1
Verbatim, from the transcripts: the passages where ImageNet comes up
The Biggest AI Deployment Nobody Talks About | Samsara CEO Sanjit Biswas
- ▶ 33:03 Sanjit Biswas That's basically from the image net era of like, okay, can we detect a mobile phone?
Cerebras CEO: Why GPUs Can't Do Fast Inference
- ▶ 26:25 Matt Turck i guess it was like four years after image net deep learning was a thing it was all it was all vain
The GPU Myth: State of AI Compute 2026 | Stephen Balaban
- ▶ 45:19 Matt Turck And that was inspired by the image net, uh, 20 and 12 moment, or that was even before that the image net moment, you know, I, I pulled the CUDA covenant repo off of Google code. 2 times in the scene
Benedict Evans: OpenAI’s Moat Problem & the Future of Software
- ▶ 58:41 Benedict Evans That's what starts working with ImageNet.
Voice AI’s Big Moment: Top Researcher on Why Everything Is Changing (Neil Zeghidour, Gradium AI)
- ▶ 23:09 Neil Zeghidour It was also obvious for me that, and for this, I think I agree 100% with Yann Lequin, the fact that what made AI dynamic and get from ImageNet in 2012 to where we are right now today,
AGI, The Future of AI Agents And The Next Wave of Opportunities in AI | Richard Socher, CEO, You.com
- ▶ 13:45 Richard Socher Um, and yeah, I was one of the coauthors of ImageNet too.
How Airtable is Ushering 500,000 Organizations Into The Era of AI | Howie Liu, CEO of Airtable
Is AI a platform shift or a paradigm shift? With Benedict Evans
- ▶ 1:17 Benedict Evans Yeah, I actually think it's, it's, it's interesting to draw a direct comparison with the last wave of machine learning, which starts in, say, sort of, 2013, 2014, so kind of a decade ago, in that it starts with demos of image recognition,…
Build and Deploy AI with Pytorch | Lightning AI Founder & CEO, William Falcon
- ▶ 2:01 William Falcon So we, like, measured neural activity in some animals, and then we used a model to take that neural activity and reconstruct, for us it was ImageNet, so images.
Make AI Less Mysterious // Serkan Piantino, Spell (FirstMark's Data Driven)
- ▶ 17:10 Serkan Piantino Um, this is the, a chart of error on the ImageNet contest, which hopefully many of you have heard of.
A Process for Discovery // Hilary Mason, Fast Forward Labs [FirstMark's Data Driven]
- ▶ 17:03 Hilary Mason I spent a lot of time in the New York City subway system, which is, I take pictures in the subway system, and there were, was no training data in the ImageNet data set of the New York City subway system.
Artificial Intelligence at Facebook // Antoine Bordes, Facebook [FirstMark's Data Driven]
- ▶ 5:59 Antoine Bordes This is basically the, the ImageNet challenge. 2 times in the scene
Richard Socher, MetaMind // Deep Learning for Enterprise (Hosted by FirstMark Capital)
- ▶ 23:38 Richard Socher We just essentially take all of ImageNet, this is a large academic data set of, with tens of thousands of different categories, and essentially subsample from all of those and say, ah, for every, every training iteration, we will randomly…
David Luan, Dextro // Real-World Video Understanding (FirstMark / Data Driven NYC)
- ▶ 9:56 David Luan So for example, a lot of the large scale visual classification work that was done, ah, and a lot of landmark results were on ImageNet, which were basically created by going to Google Image Search, ah, and searching for, for example, laptop…
Yann Lecun, Facebook // Artificial Intelligence // Data Driven #32 (Hosted by FirstMark Capital)
- ▶ 34:33 Yann LeCun Uh, you know, this open source software you can download, you can train your own convolutional net on ImageNet, and you get a pretty good recognition performance.
Mike Abbott, KPCB // Data Driven #30 // Oct 2014 (Hosted by FirstMark Capital)
- ▶ 5:45 Mike Abbott It's, it's trained on the ImageNet photos from 2010, fifteen million images, and it's actually not bad.