@Ecarter_ai Totally agree, vague prompts get vague answers. I started treating it like briefing a colleague, with the goal and the constraints up front, and the outputs got noticeably more useful.
@theo The easier to work with part is the real story. Benchmarks flip every month, but a model that gets the intent right on the first try saves more hours than any leaderboard jump.
@_philschmid Splitting the models by job instead of by hype is the sensible way to use them, Neon as the daily driver and Hydrogen for the cheap fast stuff. The three dollar versus one fifty gap matters a lot once you start running agents all day.
@dani_avila7 The enterprise deployment part is the actually useful bit, most courses stop at prompting and never touch pilots, security reviews, or how you roll it out without breaking existing workflows.
@Prathkum Three launches in a week is impressive on paper, but the naming is getting hard to follow. I now have to check which model fits the job instead of just opening Gemini and getting on with it.
@udaykotak Five years from nonexistence to a $1.5T valuation is the part that should worry every other market. It is not just Anthropic being overvalued or India being undervalued, it is the AI cycle compressing decades of value creation into a few years.
@thsottiaux Sol has been solid for my long writing sessions — it holds the thread well and rarely loses the brief halfway through. It still over-clarifies simple instructions though, so I end up repeating myself more than I should.