Kevin Wang

Everything Kevin Wang said on any show that made the record, most notable first. Each card names its show and opens the statement there.

LATENT SPACE Insight

Kevin Wang: Cross-entropy trajectory classification enables scalable deep reinforcement learning

“I think it's because we're fundamentally shifting the burden of learning from something like, Q-learning or, like, regressing to, like, TD errors, which we know is quite spurious and noisy and biased, to fundamentally, like, a classification problem. We're try…”

Kevin Wang Dec 31, 2025 ▶ 9:56 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

LATENT SPACE Assertion Supported

Wang: Deep Self-Supervised Method Beats SOTA on Goal-Conditioned RL Significantly

“We do achieve state-of-the-art performance on goal-conditioned RL and Jack's CCRL by a significant amount.”

Kevin Wang Dec 31, 2025 ▶ 21:24 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

LATENT SPACE Assertion Supported

Kevin Wang: Scaling RL network depth unlocks effective batch size scaling

“We notice that we see that scaling width actually also improves performance, and we also find that actually by scaling depth, we actually unlock the ability to scale along batch size as well.”

Kevin Wang Dec 31, 2025 ▶ 22:38 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

LATENT SPACE Assertion Supported

Kevin Wang: 1000-layer RL networks can train on single 80GB H100

“The nice thing is that all of our experiments, even the thousand layer networks, can be run on one single, 80 gigabyte, each 100 GPU.”

Kevin Wang Dec 31, 2025 ▶ 24:19 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

IN DEPTH Opinion

Wang: AI in 2025 is earlier in evolution than early-2010s mobile

“I think there's some interesting parallels, say, to where like AI is today, where I think that AI is probably a few quarters earlier than that, and people still don't know what's, what's going to happen, and there's a lot more change going on, so it's probably…”

Kevin Wang Feb 6, 2025 ▶ 36:49 Inside Braze’s blitz to $500M in CARR | Building broad, going global, and outfoxing the competition

IN DEPTH Insight

Wang: Thinking about competition under $5M ARR is generally distracting

“And even I would say up through maybe two million to five million ARR, I think that thinking too much about competition is generally distracting for two reasons. One is that you don't really know the identity of your company and also you don't know You are not…”

Kevin Wang Feb 6, 2025 ▶ 1:14:15 Inside Braze’s blitz to $500M in CARR | Building broad, going global, and outfoxing the competition

LATENT SPACE Assertion Not checkable as stated

Wang: Traditional value-based reinforcement learning fails to scale

“And so what we did is that we know that traditional RL, like let's say like value value-based RL doesn't really scale, right? This is pretty clear from the literature.”

Kevin Wang Dec 31, 2025 ▶ 4:29 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

LATENT SPACE Assertion Supported

Wang: 64 layers saturate performance in most reinforcement learning tasks

“Within our paper, like, for most environments we are able to, like, saturate, like, get to, like, almost perfect performance within just, you know, we don't even need to get to, like, a thousand layers. Like, maybe just 64 layers, for example, is sufficient.”

Kevin Wang Dec 31, 2025 ▶ 15:25 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

IN DEPTH Opinion

Wang: Early startup peers were way smarter than big consulting colleagues

“This is very, very different from big consulting, I would say. The people are way smarter, and this is a very much more scrappy environment.”

Kevin Wang Feb 6, 2025 ▶ 27:46 Inside Braze’s blitz to $500M in CARR | Building broad, going global, and outfoxing the competition

IN DEPTH Insight

Wang: Marketing creates power-law outcomes and intense competitive selection

“These marketing buyers did have, there are a lot of network effects of marketing, because if I'm running much better marketing than you are, and we have a very similar sort of brand, I'm going to win and I'm going to win really big. I'm going to win sort of li…”

Kevin Wang Feb 6, 2025 ▶ 57:40 Inside Braze’s blitz to $500M in CARR | Building broad, going global, and outfoxing the competition

LATENT SPACE Assertion Supported

Wang: GPU environments collect hundreds of millions of RL timesteps hourly

“With these, like, GPU accelerated environments, we can collect hundreds of millions of time steps of data within just a few hours”

Kevin Wang Dec 31, 2025 ▶ 17:43 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

The other half of the tape: Kevin Wang's own voice is left out of every number here. 1 statement on the record names them. every mention, with the transcript →

Statements about Kevin Wang, by other people (1)

LATENT SPACE Assertion Supported

Eysenbach: Scaling RL depth requires combining depth with residual connections

“And if we just made the depth bigger, it makes it worse. If we just add residual connections, it didn't make it better. And it was really this combination of factors that Kevin and Ishan figured out that really made this work.”

Benjamin Eysenbach Dec 31, 2025 ▶ 5:56 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

One line per show, most statements first. The link opens Kevin's full record on that show: the calibration, argument clarity, speaking style and every statement made there.

Show	Role there	Eps	Statements	Record
LATENT SPACELEDGER	Chief Product Officer, Braze	1	7	100% 5/5	full record on Latent Space →
IN DEPTHLEDGER	Chief Product Officer, Braze	1	4	—	full record on In Depth →