@CornellBaseball Despite recording 19 hits, including a five-hit performance by sophomore catcher Mason Barela, @CornellBaseball drops the series finale to Fordham, 16-11, as its ninth-inning, two-out rally fell short, leaving the tying run at the on-deck circle.
#YellCornell
Sophomore @CornellBaseball catcher Mason Barela prevents a 1-2-3 inning to begin the game, lasering a solo home run to left field to put the Big Red ahead of Richmond, 1-0, at Pitt Field.
📺: https://t.co/sjxBtFQofQ
#YellCornell
@headinthebox The real timesink is maintaining all the prompt engineering tricks that stop working 3 months later when the next model just handles it natively.
One UX challenge in GenAI systems isn’t the model itself — it’s how users experience the data behind it.
With data from multiple sources or on different schedules, late-arriving fields or partial records are inevitable.
--> The key is to design for transparency
Jeff Bezos explains the idea of “paper cut” teams
“There are big things that are really important to manage — and by the way, it’s astonishingly hard to focus on just the big things. Even though they’re obvious, they’re really hard to focus on. But in addition to that, there are all these tiny customer deficiencies. We call those ‘paper cuts,’ and we make long lists of them. Then we have dedicated teams that go fix paper cuts. That’s because the teams that are working on the big issues never get to the paper cuts. They never work their way down the list. They’re working on big things — as they should and as you want them to — so you need special teams who are charged with fixing paper cuts.”
Video source: @lexfridman (2023)
Jumping ahead early!
Freshman catcher Mason Barela capitalizes on a leadoff walk drawn by sophomore shortstop Kevin Hager, registering his second home run of the season to put @CornellBaseball ahead 2-0 in the second inning.
📺: https://t.co/PeP8faHaPK
#YellCornell
Mason Barela's first collegiate home run, a towering shot down the left-field line, has put @CornellBaseball on the board in the seventh inning, reducing its deficit against Harvard to 5-2.
📺: https://t.co/G2bjXLVHfb
#YellCornell
@andrewchen Also lots of assumptions that companies will let AI generated outputs “pass” without human validation
Does not take into account industries w strict data governance and regulations
@andrewchen From building GenAi apps for contact centers- I see it augmenting not replacing humans.
Lots of need for human in the loop to validate AI summaries, insights, etc
Shift from managing inputs raw data manipulation to managing AI generated outputs
@johncutlefish Hmmm there’s already hundreds of dashboards and millions of data points readily available but rarely used. What exact and specific use cases are you talking about and how frequently do middle managers perform them?
I suspect in future everyone is going to need to be a specialist with deep skill in something niche, as most of the generalist stuff will be done by an AI.