🚀Tech Founder & CEO @ https://nitter.cf/t.co/gn39beq4oJ 📊 | Customer & CLV 🤖Data Nerd | 💍Husband of 1 & 👨👩👧👦 Dad of 2 | 🧑💻 hacker | Tweets about DTC, Data & AI
Los Angeles, CA
Joined July 2012
- Tweets765
- Following85
- Followers431
- Likes283
Pinned Tweet
1/5 Exciting news! I'm launching a new venture, AI Whiteboard 💡, a newsletter dedicated to helping CEOs and business leaders understand the impact of AI on their organizations. #AIWhiteboard
Emad retweeted
Ever ask your AI agent to do something and it says “I don’t have access”? 🙃
Today we launched Ambiguous Workspace: 18 productivity apps reimagined so work flows between your team and agents.
Each agent gets its own identity.
Free for up to 5 teammates. Create a workspace ↓
Real example, this post: I hadn't posted on LinkedIn in 3 months. So I had Claude Code use a Playwright skill to log into my profile, scrape my recent posts, and learn what engagement patterns work for me. Then it helped me draft. From my phone. By voice.
1
56
What the fax?! I guess faxing isn’t dead 🤷♂️ Use Dropbox Fax by @DropboxSign to send and receive faxes from anywhere. #DropboxFax hellofax.com
100
Emad retweeted
Just launched Vision on @marblismAI. It takes a screenshot or your figma mockup and replicates the design system throughout your app.
The coolest part? It's not a pixel-perfect raw CSS implementation—it updates the whole design system's color scheme and attributes across your app.
Emad retweeted
In the words of Demis Hassabis: Overhyped, but underestimated:
The Generative AI landscape for April 2024
A summary of webinar with academic experts at UCLA Faculty. Featuring Andrea Eisfeldt, Gregor Schubert, Emad Hasan, David Hefter, and moderated by Lori Santikian.
open.substack.com/pub/aiwhit…
73
Brad Ito created this valuable interactive map for LLM evolution. Key takeaways:
- There's two main branches of LLMs: decoder-only and encoder-decoder model architectures
- Google / Google Brain created this world, and it really is theirs to lose.
- The latest models are not bothering with academic papers, and are using training data as a primary differentiator.
- Models free for commercial use (bold circles) continue to be made, mostly by companies hoping to profit from services for model training or inference.
bradito.me/blog/llm-map/
1
147
*Regulation of AI starting to become a growing concern*
In conversations with Gen AI builders & execs, the concerns about data have grown. Most companies now do not want to send data over the wire and are generally nervous about what the future might entail if they want to use these models.
In this video, I share my thinking about how to take a conservative approach and where I think the law and norms might land.
1
89