Video yükleniyor...
Video Yüklenemedi
Most enterprise software requires customers to write a lot of custom "glue code" — ETL pipelines, integrations, and custom workflows. By generating custom code for uncommon, company-specific use cases, LLMs have the potential to remove the need for glue code altogether. Pete Koomen
39,304 görüntüleme • 2 yıl önce •via X (Twitter)
9 Yorum

@koomen working on it, broader patterns: 1. Workflow - tools enabling you to use AI with no code @copy_ai 2. Call transcript mining - everything is recorded, now it’s data @BuildBetterAI 3. Agents for messaging - call center, SDRs Convergence of the above could eliminate ETL

@ycombinator @koomen @copy_ai Appreciate the shout out! We got some more crazy “transcript mining” comin’ out this quarter :)

@koomen @StradaHQ

@garrytan @koomen @jboursiquot

Yes 🙌🏽 the current ecosystem is complex, extremely partner led and in silos… an LLM solution that automates integration and simplifies conf. with natural language can be a game changer but the challenge will be to achieving accurate and reliable integrations, scalability and performance optimization - but focused pilot run would be a great place to start with. I am IN if someone wants to build this 🤩

@koomen That's exactly what @naologicnocode does for many customers

@koomen 7 y x x xx 4 6 xxl z6 6b. Can x y 6 ct xy ye you 5z. v g SZETSVX CAN

@koomen somebody say enterprise glue? 👋

@koomen Agreed 100%, it's glue and/or custom apps, but potential for Ai generated code to efficiently solve company-specific tasks is big - see use case #3 in article I wrote:


