More from @tyler_agg

Oct 9

how to build an AI automation (a step by step breakdown):

step 1: map out the manual task

before automating anything, document how you do it manually:

• what’s the step-by-step process if a human were to do this?
• what data and information do you need at each step?
• what are the key decision points?

step 2: decide how much variance you want in outputs

this depends on the task type:

creative tasks (writing copy) = more variance allowed

predictable tasks (categorization, data entry) = less variance needed

as this determines what guardrails you’ll need later on

Read 9 tweets

Oct 8

how to automatically scrape data from the internet (like a data engineer):

this is setting up systems that save information from the internet into organized databases

example: collecting TikTok videos, captions, and engagement metrics every day

and this data becomes the foundation for AI systems you build later on

the process for the tiktok example:

• tools like ScrapeCreators and Apify visit websites and extract the specific info you want
• saves everything to spreadsheets or databases automatically
• runs on whatever schedule you set (daily, weekly, hourly)

Read 9 tweets

Oct 7

there’s so much content on how to build AI agents, but no one ever talks about the data engineering pipelines that support them

here’s a thread going over the basics of data engineering:

the goal of data engineering:

• extract data from various sources
• transform it into structured format
• load into a data warehouse like Snowflake

and this structured data often is used context for AI systems to make personalized recommendations

why this matters for AI:

to build personalized AI systems, you need clean, structured data

as the more structured and labeled it is, the more granular/accurate context we can retrieve for an AI system

Read 10 tweets

Oct 2

how to reverse engineer any successful AI product:

step 1: understand the manual process

before diving into a technical analysis, figure out what human task this AI product is automating

> what would someone do manually to achieve the same result? > what decisions need to be made? > what data is required at each step? > what is the most painful part of this task that people are paying to automate?

step 2: create your own technical hypothesis

based on your knowledge of AI fundamentals (embeddings, RAG, APIs, etc.)

sketch out how YOU would build this

don’t overthink it - focus on the core workflow and data flow

this gives you a baseline to compare against

Read 8 tweets

Oct 1

how to discover AI business ideas that actually make money (step by step breakdown):

step 1: research industries with manual bottlenecks

use this prompt to understand where people are struggling:

“You’re a 20-year veteran in the [INDUSTRY]. What tasks consume the most time daily? What repetitive work do you wish would disappear?”

test this across different sectors

step 2: reach out to real people in those industries

LinkedIn search for decision-makers and send this:

“I’m researching the [INDUSTRY] and would love insights from someone with experience.

Would you be interested in a quick call to ask about your daily daily operations? Happy to send $50 for your time“

(if you’re in college you can just say it’s for a school project lol)

Read 7 tweets

Sep 30

how to vibe code your AI project in 7 days (step by step breakdown):

day 1: map out your system design

don’t jump straight into code - plan the complete workflow first

what steps does your tool need to get users from point A to point B?

create a document outlining each step, what data is needed, what decisions get made

note that this isn’t about code or AI yet - it’s pure logic

“step 1: user uploads document, step 2: extract key info, step 3: if condition X then do Y”

having a clear system design prevents building something that doesn’t work and makes it much easier to build

Read 11 tweets