Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Building an AI Agent for 11k Users
Search
Phil Calçado
September 03, 2024
59
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Building an AI Agent for 11k Users
Phil Calçado
September 03, 2024
More Decks by Phil Calçado
See All by Phil Calçado
RubyConf Brasil 2020: Rails não escala. (mas você não precisa de escalabilidade)
philcalcado
0
200
Why code in Node.js often gets rejected by SoundCloud
philcalcado
2
180
Featured
See All Featured
Mobile First: as difficult as doing things right
swwweet
225
10k
KATA
mclloyd
PRO
35
15k
Designing Experiences People Love
moore
143
24k
Stewardship and Sustainability of Urban and Community Forests
pwiseman
0
230
Navigating Team Friction
lara
192
16k
Information Architects: The Missing Link in Design Systems
soysaucechin
0
970
How to build an LLM SEO readiness audit: a practical framework
nmsamuel
1
780
Visual Storytelling: How to be a Superhuman Communicator
reverentgeek
2
560
Measuring Dark Social's Impact On Conversion and Attribution
stephenakadiri
2
220
WCS-LA-2024
lcolladotor
0
630
A Modern Web Designer's Workflow
chriscoyier
698
190k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
47
8.2k
Transcript
None
Outropy is the PaaS that enables engineers to build Compound
AI apps and Agents… check us out https://outropy.ai
…but Outropy was the vscode for everything you do that
isn’t writing code
• Cache invalidation • Naming things • “Book me a
meeting every monday with everyone from the web team” We have 5 mins, let’s pick a hard problem to go over
Attempt #1- What we now call RAG-Fusion
ERR_LLMS_LOVE_MAKING_SHIT_UP
Attempt #2- What we now call GraphRAG
just because I am in this channel/ commit to this
repo/comment on this Jira doesn’t mean I am part of this team ERR_ENTERPRISES_ARE_CHAOTIC
Attempt #3- The Probablility Enterprise Graph
can’t cross train with customer data and even the largest
customers don’t have enough data to make the model work ERR_NOT_ENOUGH_DATA
Attempt #4- Don’t boil the ocean
ERR_ actually, this one wasn’t too bad • 11,000 Users
• ~40 Companies • 10x better than Slack* AI for 1/3 of the price • I miss it every day 😢 * In our own becnhmarks (we are still waiting to hear back from MKBHD)
tl;dr you are not here to solve AGI keep the
main thing the main thing
None