Graph RAG Works Better Than Standard RAG
GraphRAG leverages structural information across entities to enable more precise and comprehensive retrieval, capturing relational knowledge and facilitating more accurate, context-aware responses.
This improves the accuracy of standard RAG systems.
Abacus AI Enterprise autonomously builds the best RAG system with big data
Micro Agent: an AI tool that generates unit tests and iterates on code until the tests pass, ensuring reliable function creation in languages like JavaScript, TypeScript, and Python
My top models for various use cases
Video - Hailuo and Kling
Images - Flux Pro, Grok and MJ
Code - Sonnet, o1
Writing - GPT-4o, Sonnet
Video analysis- Gemini
Low latency - GPT-4o mini, Flash 2.0
Data analysis - o1
COT - o1
RAG - 4o, Sonnet
Audio - ElevenLabs
AI is continuing to improve on all of these use cases and we will see a step change by the end of 2025
I built an AI Data Analysis Agent.
Just upload CSV/Excel files and analyze data through natural language. It uses DuckDB for lightning-fast processing.
100% Opensource Code with step-by-step tutorial.
35 STARTUP IDEAS TO START IN 2025 (saas, ai agents etc)
1. AI agent that turns customer testimonials into multiple formats - social proof, case studies, sales decks. marketing teams need this daily. $300/month.
2. agent that turns product demo calls into instant microsites. sales teams record hundreds of calls but waste the content. $200 per site, scales to thousands.
3. fitness AI that builds perfect workouts by watching your form through phone camera. adjusts in real-time like a personal trainer. $30/month
4. directory of enterprise AI budgets and buying cycles. sellers need signals. charge $1k/month for qualified leads.
5. AI detecting wasted compute across cloud providers. companies overspending $100k/year. charge 20% of savings. win-win
6. tool turning customer support chats into custom AI agents. companies waste $50k/month answering same questions. one agent saves 80% of support costs.
7. agent monitoring competitor API changes and costs. product teams missing price hikes. $2k/month per company.
8. tool finding abandoned AI/saas side projects under $100k ARR. acquirers want cheap assets. charge for deal flow. Could also buy some of these yourself. Build media business around it.
9. AI turning sales calls into beautiful microsites. teams recreating same demos. saves 20 hours per rep weekly.
10. marketplace for AI implementation specialists. startups need fast deployment. 20% placement fee.
11. agent streamlining multi-AI workflow approvals. teams losing track of spending. $1k/month per team.
12. marketplace for custom AI prompt libraries. companies redoing same work. platform makes $25k/month.
13. tool detecting AI security compliance gaps. companies missing risks. charge per audit.
14. AI turning product feedback into feature specs. PMs misinterpreting user needs. $2k/month per team.
15. agent monitoring when teams duplicate workflows across tools. companies running same process in Notion, Linear, and Asana. $2k/month to consolidate.
16. agent converting YouTube tutorials into interactive courses. creators leaving money on table. charge per conversion or split revenue with them.
17. marketplace for AI-ready datasets by industry. companies starting from scratch. 25% platform fee.
18. tool finding duplicate AI spend across departments. enterprises wasting $200k/year. charge % of savings.
19. AI analyzing GitHub repos for acquisition signals. investors need early deals. $5k/month per fund.
20. directory of companies still using legacy chatbots. sellers need upgrade targets. charge for leads.
21. agent turning Figma files into full webapps. designers need quick deploys. charge per site. Could eventually get acquired by framer or something.
22. marketplace for AI model evaluators. companies need bias checks. platform makes $20k/month.
23. tool detecting AI policy violations in comms. legal teams missing risks. $2k/month per company.
24. AI turning business metrics into investor updates. founders waste days monthly. $500 per update.
25. tool detecting when APIs silently change specs. engineering teams waste weeks debugging. $1k/month per integration monitored.
26. tool detecting when competitors accidentally leak roadmap details in meta tags, GitHub commits, and job posts. product teams need signal. $2k/month per competitor tracked.
27. marketplace for AI training video creation. companies need custom data. charge per project.
28. platform finding underpriced acquisition targets by scanning for companies with high NPS but poor marketing. $10k/month to investors.
29. marketplace for sharing expensive API credits. companies overbuy capacity they don't use. take 15% of trades.
30. agent converting PDFs into searchable knowledge bases. teams losing docs. $1k/month
31. AI turning bug reports into test cases. devs waste hours rewriting. charge per month.
32. AI that turns your photos into perfect social posts with your writing style. end Instagram anxiety. $15/month.
33. AI turning your Kindle highlights into personalized learning materials. remember what you read. $20/month.
34. agent turning company slack/notion into personalized onboarding. teams waste weeks training new hires. $5k/month per company
35. marketplace for AI training data from failed startups. acquirers need datasets. platform makes $30k/month in deals.
build something people need. make it simple. charge money. don’t bookmark this, build it. im rooting for you
good luck.
All of the past issues of the Huberman Lab Neural Network Newsletter are accessible at zero cost (no sign-up required).
Here are the most popular newsletters:
Chart lovers check $DOGE price action Dec 2020 to April 2021
Use higher timeframes only for comparison
Now open $COCOS & compare the two
What if $COCOS can pull $DOGE in 2022?
Possible subject to #Bitcoin PA
Looking for COINS that can do 10-100x in 2022, comment your picks
@Hu_barts @devchart Full stack and mobile development experience, plus some product management here and there.. I fit in a lot of positions, so depends on what's available really :)
@dcfgod actually this was my first play and if you don't mind explaining, what was the game supposed to be? if you insta sell you win? else you lose? how did the price jump up so high then dumped so fast?
Noticed lots of smart new blood in crypto missing out on the good ol’ wisdom.
Below you can find a weekend reading of what I think are the “post-2016 crypto classics”: