@uzair__afridi Reverse engineering "what is correct" is an interesting problem to solve. Rerank has set a paradigm that is not big enough. We need sequential thinking in rerank to better determine what is valid and contextually correct
@berman66 Have you use ServiceNows platform since 2024??? Not good for them to build agents, they need to fix their core product and make it more agent friendly, focusing on teaching their customers how to adopt AI rather than buy AI poorly laid over products thatve gotten worse over time
@wholemars I make everyone try the brisket when they come over. Can you imagine paying for BRISKET for the ENTIRE COMPANY and people skipping it? Id be livid
@stevekrouse This is exactly our thesis at https://t.co/ifBmUoFBrE (site is outdated). We're testing this model with consulting and engineering agencies now.
With Ocho you talk with clients, internally, etc. and the rest happens without you. You only do the part that matters: creative work
@JakobASimmons January 26 it was clear that we were heading to a world of fractionalized market share in SaaS, like YouTube did to Hollywood. I thought it would be due to good new apps, I didn’t think that all the big targets would be taken down
@GIANFRANCOGV Claude is faster and better at design than the “Design” product. Claude design felt more like a kids toy with decorative controls. Nobody would ever use them if they can just talk and all the clicking goes away
@humphreybc My take: Dario said slow the model development, so they pointed tokens at some side quests. To anthropic, replacing 30 years of technology is a side quest. I’d imagine they also harvest better data on their own tools
@ernietedeschi this is great for startups & corps under 100 employees IMO, but will not touch enterprise. I work with enterprises still struggling to adopt MS teams and do everything on email
@jackmcclelland i asked chatgpt "what's good baby name for a girl in 2026" before we wrote her name on the birth certificate to make sure we don't end up lumped in with "people" that let google ai summaries make up their minds