Listening to your most devoted users can lead you wrong because they have their own specific needs, which may not represent the needs of the wider market. In fact, they don't want you to solve for the mass market – the classic "Chasm" for new products. I just saw this on Steam.
AI
-
Jensen Huang: AI Makes Everyone a Programmer
By
–
Makes me think about what Jensen Huang said a couple weeks ago.
— Rowan Cheung (@rowancheung) 12 mars 2024
"Everybody in the world is now a programmer"https://t.co/x0Jhqc2v6uMakes me think about what Jensen Huang said a couple weeks ago. "Everybody in the world is now a programmer"
-
AI Agents and Prompt Engineering Transform Developer Productivity
By
–
If devs using AI are already working 4x faster, I wonder what happens when they start deploying AI software engineers alongside them.
— Rowan Cheung (@rowancheung) 12 mars 2024
Feels like everyone will be a 10x dev if you know how to prompt LLMs/Agents.
And the 10x devs now will be 100x devs pic.twitter.com/vqctjWEPKsIf devs using AI are already working 4x faster, I wonder what happens when they start deploying AI software engineers alongside them. Feels like everyone will be a 10x dev if you know how to prompt LLMs/Agents. And the 10x devs now will be 100x devs
-
Devin AI Autonomously Completes Full Upwork Job
By
–
This demo is also wild.
— Rowan Cheung (@rowancheung) 12 mars 2024
They gave Devin a job posted in Upwork, and it autonomously completed the whole thing.pic.twitter.com/OKG7T2eThFThis demo is also wild. They gave Devin a job posted in Upwork, and it autonomously completed the whole thing.
-
Devin: Cognition AI’s Autonomous Coding Agent Resolves Issues
By
–
So, Cognition AI just introduced Devin, the most capable autonomous coding agent I've seen.
— Rowan Cheung (@rowancheung) 12 mars 2024
Instead of just autocompleting tasks, it can write entire apps on its own in minutes.
It also correctly resolves 13.86% of issues, autonomously. pic.twitter.com/7EWncjRQmdSo, Cognition AI just introduced Devin, the most capable autonomous coding agent I've seen. Instead of just autocompleting tasks, it can write entire apps on its own in minutes. It also correctly resolves 13.86% of issues, autonomously.
-
Meta’s 24k H100 Cluster Pods Infrastructure for Llama3
By
–
Here's details on Meta's 24k H100 Cluster Pods that we use for Llama3 training.
* Network: two versions RoCEv2 or Infiniband. * Llama3 trains on RoCEv2
* Storage: NFS/FUSE based on Tectonic/Hammerspace
* Stock PyTorch: no real modifications that aren't upstreamed
* NCCL with -
Scaling Learning Rate with Batch Size in Model Training
By
–
I think it really depends on the batch size you are using. Let's say you are multiplying the batch size by x, then you should multiply the learning rate by sqrt(x)
-
HRI Conference Workshop on Embodied Voices and TTS Models
By
–
Audience question on character voices and TTS needs @HRI_Conference – Consider our newly accepted RSS Workshop on Embodied Voices on July 15! Bringing the latest deep TTS models together with roboticists https://
rosielab.github.io/wev/ #hri2024 -

Dr. Kennedy’s Tips for Designing Robot Characters and Holographic Personas
By
–
Dr. James Kennedy opens up the 2024 @HRI_Conference
, discussing robots and holographic characters. His 4 tips: design characters (not agents), create coherent and consistent personality, allow for groups, and make people feel heard. #hri2024 -
Self-driving predictions continue to miss their timelines
By
–
So far the timing on every prediction about self-driving being around the corner has turned out to be wrong. See my annual dated predictions and updates at http://
rodneybrooks.com/blog for details. I appreciate this update from @IEEESpectrum
.
