Muse and Instinct are the first credible threats to ChatGPT since launch.
It is accessible. It is everywhere consumers already are.
The model performance is great for consumer use cases.
Bad for OpenAI.
Amazon has blocked Meta Platforms’s new artificial intelligence agent from its retail site after the social media company declined a request to remove the bot https://t.co/XlzL1HLL5n
Will enterprises keep paying a premium for proprietary models or shift to open source? Who knows! It’ll probably come down to whether frontier labs can build capabilities open source rivals can’t replicate.
But whichever way it resolves, tokens still run through a hyperscaler.
Google always wins.
Why bother fighting for the best model, when you own the engine that powers it all.
Ranking Claude by task, god to intern:
1. Code
2. Excel
3. App design
4. Slides
….
100. Writing a document I'd actually send 😭
Time to relabel large “language” models as large “math” models?
Managing agents and people appears similar, but isn’t. Both need constant context switching across parallel work. But with people quality is guaranteed by a whole system - hiring, incentives, accountability. With agents you pay that cost for variable output. Not always efficient!
Started using Glean recently - what a product! Connects to everything across email, drive, notes. Completely removes the friction of uploads, data limits, and detailed context prompts.
Happy to let the agents work while I doomscroll in peace.
Lost my nana recently, and reflecting on his life - particularly how his passion for his work (lawyer) helped him cope with the loss of two daughters. He was an inspiration, who taught me that work can truly be a source meaning, direction, and happiness
@AkshatAg18 Another way to do it would be to just have shorter papers, which are actually readable. It would also teach you skills in writing, synthesis, communication that are actually of use.