Nahid Alam
@nahidalam
Building VLMs @meraki. Community @Cohere_Labs Founded @litehouseio Alum @backstage_cap, @intel. #TEDx. appeared on #startup showcase @TBSNetwork
Finetuned @nvidia GR00T N1 for bimodal robot arm manipulation and won the 1st prize at @LeRobotHF hackathon. Thanks to my amazing collaborators 🙏
We got lots of great community feedback on our open-source GR00T N1! Check out our Github, star, fork, contribute back! Let's solve generally intelligent robots together, one commit at a time. github.com/NVIDIA/Isaac-G…
Good summary. I would add execution to it though. Not everyone “sees”. Not everyone who “sees” can execute
i like thinking of it as a combination of ... seeing applicable patterns between different fields, developing intuition as a competency, and having experiences that drive a point of view.
The opportunity gap in AI is more striking than ever. We talk way too much about those receiving $100M or whatever for their jobs, but not enough those asking for <$1k to present their work. For 3rd year in a row, @ml_collective is raising funds to support @DeepIndaba attendees.
If you want to make the wrong decision: Ask everyone for their opinion, and ignore your gut.
Anyone knows why they did not compare CLIP in Figure 4 of SigLIP2 paper? cc: @giffmana I am seeing LLaVA-SigLIP2 performs worse in TallyQA compared to vanilla LLaVA (that used CLIP). Trying to root cause.
No, 94066 is NOT San Carlos, CA. And I didn't even want to know the city name.
😂
it’s really a miracle how humanity accomplished anything at all before we had stakeholder alignment and could properly frame the context
If you don't believe you are better than most people at *something,* you'll never do more than most people do. Better at a skill ≠ Superior human Most people can be exceptional at *something* -lean into it. Don't hold yourself to mediocre because you're afraid to be great.
The smarter your AI is, the less data it needs. Needing huge training sets is a sign of stupidity.
When you just decide to stop hiding, hesitating or playing small - a funny thing happens... The world adjusts. Life catches up - it will filter some out, you'll gain others, and real fantastic living begins.
Most success stories look a lot like ordinary effort repeated to an unreasonable degree.
i don't really understand the people who are worried about AI taking jobs. jobs suck! i never want to work at a job another day in my life
Deep. If it is not, change your group
Your group chat should be more interesting than your timeline.
Most things worth waiting for, are also worth working for, which you don’t need to wait to start working for.
😂
Management: You know what the world really needs? Devs: ? Management: A new vscode fork. Devs: ??? Management: OK get to work.
It is cool to be capable. It is cool to know shit. That's why the HF team is open-sourcing not just the model, but the training code and datasets too. Learn. Build. Make it your own. github.com/huggingface/sm…
Amen
One of the best parts of leverage is the ability to disconnect from things you hate. - Don’t like a client? Fire them. - Dislike part of your business? Close it. - Losing interest in a direction? Change it. There’s so much freedom in saying “no thanks” more often.
One of the first technical reports in a while that has considerable detail. In particular -- preference training tends to be much more painful for multimodal, largely because of lack of calibrated rewards. Some nice details here about overcoming some of those challenges.
Me: I ran code from open source LLM papers :)
interviewer: "what is the hardest technical challenge you've faced" me: "i ran code from a 3d research paper"