Drew Thomas

systems engineer → solo builder → aspiring futures trader
didn't code until my 30s. AI closed the gap.
building @noverloadapp. shipping in public

[Drew Thomas @DrewAutomates]

no video team. no screen recording.

claude code + remotion /skills + ~1 hr + a few prompts

the wild part: it pulls directly from supabase via mcp. the app literally renders itself

posting my workflow next

who else has been testing remotion + claude code?
drop your videos

[Remotion @Remotion]

Remotion now has Agent Skills - make videos just with Claude Code!

$ npx skills add remotion-dev/skills

This animation was created just by prompting 👇

[Drew Thomas @DrewAutomates]

might not look like much but this is the result of 8+months of research and going deep down the algorithmic trading rabbit hole

not a single trade executed by me today

built with claude code with research by noverload

more on this coming soon

[Drew Thomas @DrewAutomates]

nobody's getting rich letting an agent trade their robinhood account

the people getting rich are selling retail the agent

but the signal is loud: agent trading just became infrastructure

next 18 months are about one.. thing using ai to translate real edge into algos not agents

[Robinhood @RobinhoodApp]

Your strategy shouldn't sleep just because you do.

Connect your AI agent to a Robinhood Agentic Account to explore trade ideas, build and rebalance portfolios, program custom tools, and place trades as your strategy evolves.

Rolling out now.

Learn more: https://t.co/Vs3yjWkjTN

[Drew Thomas @DrewAutomates]

@kenbuildsalone @shipordie_ Dope kickoff video!

[Drew Thomas @DrewAutomates]

did not expect to wake up and become pirate

joined shipordie by @jackfriks & @marclou

the gap between "im building this" and "I shipped this" is accountability
find a crew that calls you out or keep telling yourself soon

took the plunge. looking forward to meeting the crew!

[Drew Thomas @DrewAutomates]

@eliana_jordan test out remotion skills if you haven't! game changer for creating videos and product demos https://t.co/4KHwVrfOxC

[Drew Thomas @DrewAutomates]

no video team. no screen recording.

claude code + remotion /skills + ~1 hr + a few prompts

the wild part: it pulls directly from supabase via mcp. the app literally renders itself

posting my workflow next

who else has been testing remotion + claude code?
drop your videos

[Drew Thomas @DrewAutomates]

Just hit $20MRR!

Might not seem like much but is a signal to keep building and shift to marketing

Recent Milestones
- 704 things saved
- 84 signups
- SEO bringing in 1-2 users regularly

Now the focus is on conversion and positioning

[Drew Thomas @DrewAutomates]

saas builders: fix your indexing
i ignored it for months. traffic was dead.

spent weeks fixing it. now seeing:
google traffic up → chatgpt sending visitors → reddit too

seo in 2026 isn't just google anymore.

[Drew Thomas @DrewAutomates]

karpathy second brains, knowledge bases, and personal wiki trend is real

noverload: 51 users. ~500 saves. steady growth

building easy. reaching the right people is the real game

if you're saving content and want to actually use it later - this is what i'm building

[Drew Thomas @DrewAutomates]

@alex_prompter Built it and turned it into an app
https://t.co/VOffSL8hN2

[Drew Thomas @DrewAutomates]

"I think there is room here for an incredible new product instead of a hacky collection of scripts."

agreed.

been building Noverload for exactly this

save content. compile it. query it. connect to claude via MCP

the personal knowledge base that AI can actually use
Demo here

[Drew Thomas @DrewAutomates]

@AliAbdaal @karpathy hey Ali! i've made something very similar and would love for you to check it out.

it's a web app that users drop links to generate summaries and knowledge bases. it has the ability to import from readwise as well!

here's a quick video overview
https://t.co/liqmgIY35k

[Drew Thomas @DrewAutomates]

"I think there is room here for an incredible new product instead of a hacky collection of scripts."

agreed.

been building Noverload for exactly this

save content. compile it. query it. connect to claude via MCP

the personal knowledge base that AI can actually use
Demo here

[Drew Thomas @DrewAutomates]

@joncphillips Thank you!

[Drew Thomas @DrewAutomates]

"I think there is room here for an incredible new product instead of a hacky collection of scripts."

agreed.

been building Noverload for exactly this

save content. compile it. query it. connect to claude via MCP

the personal knowledge base that AI can actually use
Demo here

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Drew Thomas @DrewAutomates]

Saw karpathy's post about LLM knowledge bases and
realized I'm we're already 80% there with noverload

shipped concept pages this week:

save sources on a topic → auto-compiles into a wiki page

going from "save for later" tool → personal knowledge compiler

[Drew Thomas @DrewAutomates]

@HilaShmuel looks cool! definitely checking out. i built something similar but for consuming content and building knowledge wikis. funny how everything happens so fast posted my demo here if interested https://t.co/liqmgIY35k

[Drew Thomas @DrewAutomates]

"I think there is room here for an incredible new product instead of a hacky collection of scripts."

agreed.

been building Noverload for exactly this

save content. compile it. query it. connect to claude via MCP

the personal knowledge base that AI can actually use
Demo here

[Drew Thomas @DrewAutomates]

@karpathy @surim0n This is what https://t.co/WtR4jOF3qV aims to solve

[Drew Thomas @DrewAutomates]

claude code is now my trade engineer

reverse engineered my strategy. coded it into sierra chart. now it trades for me

first day actually profitable so far: +$444

early, but something might be clicking

[Drew Thomas @DrewAutomates]

been heads down. here's what shipped lately:

– second agency project delivered (awards platform)
– noverload growing (44 users now)
– building an automated trading system with claude code (still losing money)

less posting. more output.

felt good. but i'm back!

[Drew Thomas @DrewAutomates]

it's alive!

everyone's talking about openclaw and polymarket trading bots

meanwhile i'm over here using claude code to trade futures

now i can lose money faster than ever before.
automation is beautiful

[Drew Thomas @DrewAutomates]

everyone who's tried automated trading has a story about why it didn't work

spent the weekend building one anyway with claude code

either ai changes the game or i join the graveyard.
let's see

[Drew Thomas @DrewAutomates]

build in public update!

→ 44 users
→ 480 content saves
→ 109 youtube, 159 tweets, 186 articles saved

not hockey stick growth. just people actually using it.

that's enough to keep going