Isaac Flath
@isaac-flath.bsky.social
R & D @ Answer.ai
Currently working on: gallery.fastht.ml | https://fh-frankenui.answer.ai | solveit.fast.ai | https://tryvirgil.com/
Currently working on: gallery.fastht.ml | https://fh-frankenui.answer.ai | solveit.fast.ai | https://tryvirgil.com/
I quit my job ~6 months ago to focus on learning and having a bigger impact.
The AI coding course with Eleanor Berger is one (of several) projects that is better than I imagined on both fronts
Cohort 1 was a huge success, and Cohort 2 in January is gonna be even better 😄
The AI coding course with Eleanor Berger is one (of several) projects that is better than I imagined on both fronts
Cohort 1 was a huge success, and Cohort 2 in January is gonna be even better 😄
December 22, 2025 at 1:54 AM
I quit my job ~6 months ago to focus on learning and having a bigger impact.
The AI coding course with Eleanor Berger is one (of several) projects that is better than I imagined on both fronts
Cohort 1 was a huge success, and Cohort 2 in January is gonna be even better 😄
The AI coding course with Eleanor Berger is one (of several) projects that is better than I imagined on both fronts
Cohort 1 was a huge success, and Cohort 2 in January is gonna be even better 😄
My favorite thing about reducing token usage for coding agents with better search with @mixedbreadai
December 12, 2025 at 4:57 PM
My favorite thing about reducing token usage for coding agents with better search with @mixedbreadai
Multi-vector search means semantic search is back for coding agents...and it was always clear it would
December 11, 2025 at 9:58 PM
Multi-vector search means semantic search is back for coding agents...and it was always clear it would
60% of tokens are spent searching and exploring the codebase.
Agents need better search.
Agents need better search.
December 11, 2025 at 4:59 PM
60% of tokens are spent searching and exploring the codebase.
Agents need better search.
Agents need better search.
mgrep helps explore complex documents and pdfs
December 10, 2025 at 10:01 PM
mgrep helps explore complex documents and pdfs
AI coding agents burn tokens guessing keywords for grep and flood the context window with noise
There's a better way.
I hosted a talk by @ruithebaker, a founding engineer at @mixedbreadai, about their solution.
mgrep 🧵
elite-ai-assisted-coding.dev/p/mgrep-wit...
There's a better way.
I hosted a talk by @ruithebaker, a founding engineer at @mixedbreadai, about their solution.
mgrep 🧵
elite-ai-assisted-coding.dev/p/mgrep-wit...
mgrep with Founding Engineer Rui Huang
The Problem with grep for AI Agents
elite-ai-assisted-coding.dev
December 10, 2025 at 6:28 PM
AI coding agents burn tokens guessing keywords for grep and flood the context window with noise
There's a better way.
I hosted a talk by @ruithebaker, a founding engineer at @mixedbreadai, about their solution.
mgrep 🧵
elite-ai-assisted-coding.dev/p/mgrep-wit...
There's a better way.
I hosted a talk by @ruithebaker, a founding engineer at @mixedbreadai, about their solution.
mgrep 🧵
elite-ai-assisted-coding.dev/p/mgrep-wit...
Q: Why is semantic search for code coming back? (Cursor, mgrep, etc)
A: Multi-vector architecture
Q: Why do I care?
MUCH less token usage + better responses if used
Q: What makes it work?
A: Token (not chunk) level embeddings with extreme quantization
Here's what I learned 🧵
A: Multi-vector architecture
Q: Why do I care?
MUCH less token usage + better responses if used
Q: What makes it work?
A: Token (not chunk) level embeddings with extreme quantization
Here's what I learned 🧵
December 1, 2025 at 7:13 PM
Q: Why is semantic search for code coming back? (Cursor, mgrep, etc)
A: Multi-vector architecture
Q: Why do I care?
MUCH less token usage + better responses if used
Q: What makes it work?
A: Token (not chunk) level embeddings with extreme quantization
Here's what I learned 🧵
A: Multi-vector architecture
Q: Why do I care?
MUCH less token usage + better responses if used
Q: What makes it work?
A: Token (not chunk) level embeddings with extreme quantization
Here's what I learned 🧵
I got better quality, half the tokens, and twice the speed with MixedBread's mgrep
The "trick" is a multi-vector architecture. Why now? Because the arch and quantization details are hard
Join Aamir Shakir (MixedBread Founder) for a research talk on how it works 👇
maven.com/p/0c0eed/mo...
The "trick" is a multi-vector architecture. Why now? Because the arch and quantization details are hard
Join Aamir Shakir (MixedBread Founder) for a research talk on how it works 👇
maven.com/p/0c0eed/mo...
Modern Multi-Vector Code Search
Mixedbread showed in their launch how much faster/better semantic search could make Claude code. Cursor also just announced semantic embedding support, and other agents are soon to follow. I got better quality with 1/2 tokens and almost 2x faster with mixedbread search.
Learn what changed from the leading researchers in the space.
maven.com
November 26, 2025 at 7:08 PM
I got better quality, half the tokens, and twice the speed with MixedBread's mgrep
The "trick" is a multi-vector architecture. Why now? Because the arch and quantization details are hard
Join Aamir Shakir (MixedBread Founder) for a research talk on how it works 👇
maven.com/p/0c0eed/mo...
The "trick" is a multi-vector architecture. Why now? Because the arch and quantization details are hard
Join Aamir Shakir (MixedBread Founder) for a research talk on how it works 👇
maven.com/p/0c0eed/mo...
RAG isn't dead, it evolved
I compared the new semantic search (RAG) against default agentic search (also RAG 😄) for coding agents. Semantic search won
Rui Huang ( MixedBread ) will give an engineering talk on why, intuition, and setup
Sign up 👇
maven.com/p/9c51af/th...
I compared the new semantic search (RAG) against default agentic search (also RAG 😄) for coding agents. Semantic search won
Rui Huang ( MixedBread ) will give an engineering talk on why, intuition, and setup
Sign up 👇
maven.com/p/9c51af/th...
The 1/2 Token Codebase Search.
RAG famously (and incorrectly) written off as dead by many when agentic search became more popular for agentic coding. Recent advances in semantic search have dramatically outperformed agentic search. This engineering talk gives you the intuition and information to be an expert user.
maven.com
November 25, 2025 at 6:25 PM
RAG isn't dead, it evolved
I compared the new semantic search (RAG) against default agentic search (also RAG 😄) for coding agents. Semantic search won
Rui Huang ( MixedBread ) will give an engineering talk on why, intuition, and setup
Sign up 👇
maven.com/p/9c51af/th...
I compared the new semantic search (RAG) against default agentic search (also RAG 😄) for coding agents. Semantic search won
Rui Huang ( MixedBread ) will give an engineering talk on why, intuition, and setup
Sign up 👇
maven.com/p/9c51af/th...
Using Alpine.js in Python frameworks like Air, Fasthtml, or Dash often means you end up writing code in strings
Writing code inside a string is a terrible dev experience and hard to maintain. But I also kinda hate jinja TBH
Writing code inside a string is a terrible dev experience and hard to maintain. But I also kinda hate jinja TBH
November 16, 2025 at 9:43 PM
Using Alpine.js in Python frameworks like Air, Fasthtml, or Dash often means you end up writing code in strings
Writing code inside a string is a terrible dev experience and hard to maintain. But I also kinda hate jinja TBH
Writing code inside a string is a terrible dev experience and hard to maintain. But I also kinda hate jinja TBH
This week I was fixing some basic and classic async problems: a "processing" pill that never updated to "complete." Blech.
Read the details from my new newsletter 👇
isaacflath.com/newsletter/2
Read the details from my new newsletter 👇
isaacflath.com/newsletter/2
Async Background Tasks - Newsletter
Hey this is Isaac,
This week I was fixing some basic and classic async problems: a
isaacflath.com
November 8, 2025 at 7:50 PM
This week I was fixing some basic and classic async problems: a "processing" pill that never updated to "complete." Blech.
Read the details from my new newsletter 👇
isaacflath.com/newsletter/2
Read the details from my new newsletter 👇
isaacflath.com/newsletter/2
Faced a classic async UI bug: a "processing" pill didn't update
I used AI to brainstorm to get 4 patterns (I chose self-polling pill)
AI is good at giving common patterns fast so I can spend more time on decisions and considering less obvious options
elite-ai-assisted-coding.dev/p/using-an-...
I used AI to brainstorm to get 4 patterns (I chose self-polling pill)
AI is good at giving common patterns fast so I can spend more time on decisions and considering less obvious options
elite-ai-assisted-coding.dev/p/using-an-...
Using an AI Assistant to Brainstorm
I ran into a classic background task UI problem in my app this week.
elite-ai-assisted-coding.dev
November 7, 2025 at 9:57 PM
Faced a classic async UI bug: a "processing" pill didn't update
I used AI to brainstorm to get 4 patterns (I chose self-polling pill)
AI is good at giving common patterns fast so I can spend more time on decisions and considering less obvious options
elite-ai-assisted-coding.dev/p/using-an-...
I used AI to brainstorm to get 4 patterns (I chose self-polling pill)
AI is good at giving common patterns fast so I can spend more time on decisions and considering less obvious options
elite-ai-assisted-coding.dev/p/using-an-...
AI demos often get stuck in "Proof of Concept Purgatory" by jumping to agents too early.
Hugo shares a framework for building AI products, by progressing through the "agentic spectrum" iteratively , starting with augmented LLMs and workflows
elite-ai-assisted-coding.dev/p/stop-buil...
Hugo shares a framework for building AI products, by progressing through the "agentic spectrum" iteratively , starting with augmented LLMs and workflows
elite-ai-assisted-coding.dev/p/stop-buil...
Stop Building AI Agents... And Start Building Smarter AI Workflows
I was excited to host Hugo in our course for a talk.
elite-ai-assisted-coding.dev
October 31, 2025 at 6:35 PM
AI demos often get stuck in "Proof of Concept Purgatory" by jumping to agents too early.
Hugo shares a framework for building AI products, by progressing through the "agentic spectrum" iteratively , starting with augmented LLMs and workflows
elite-ai-assisted-coding.dev/p/stop-buil...
Hugo shares a framework for building AI products, by progressing through the "agentic spectrum" iteratively , starting with augmented LLMs and workflows
elite-ai-assisted-coding.dev/p/stop-buil...
AI Assisted Coding Cohort 1 just ended and feedback is coming in. Really happy to hear V1 was beneficial to so many people ❤️
Feedback also means huge improvements coming for cohort 2 😁
I still have quite a lot to add, but I'm adding reviews here!
isaacflath.com/course/elit...
Feedback also means huge improvements coming for cohort 2 😁
I still have quite a lot to add, but I'm adding reviews here!
isaacflath.com/course/elit...
Elite AI Assisted Coding - Transform Your Development Workflow
Learn to build better software and learn more with LLMs and AI agents.
isaacflath.com
October 25, 2025 at 9:52 PM
AI Assisted Coding Cohort 1 just ended and feedback is coming in. Really happy to hear V1 was beneficial to so many people ❤️
Feedback also means huge improvements coming for cohort 2 😁
I still have quite a lot to add, but I'm adding reviews here!
isaacflath.com/course/elit...
Feedback also means huge improvements coming for cohort 2 😁
I still have quite a lot to add, but I'm adding reviews here!
isaacflath.com/course/elit...
Learn from Open Hands, a powerful open-source async agent platform, about how to build your own agents and how their agent SDK helps.
maven.com/p/0fe39a/op...
maven.com/p/0fe39a/op...
OpenHands: AI software development agent platform
OpenHands (and All Hands, the company behind it) set to push the envelope with an agentic software development platform, all while remaining fully committed to their open-source origins.
maven.com
October 18, 2025 at 5:37 PM
Learn from Open Hands, a powerful open-source async agent platform, about how to build your own agents and how their agent SDK helps.
maven.com/p/0fe39a/op...
maven.com/p/0fe39a/op...
RAG is how you use search to gives users what they need.
My new RAG course just launched, I hope to see you there!
My new RAG course just launched, I hope to see you there!
October 14, 2025 at 4:01 PM
RAG is how you use search to gives users what they need.
My new RAG course just launched, I hope to see you there!
My new RAG course just launched, I hope to see you there!
I love hearing feedback like this about the new RAG course ❤️
To learn more about the course go here: elite-ai-assisted-coding.dev/p/learn-rag...
To learn more about the course go here: elite-ai-assisted-coding.dev/p/learn-rag...
October 12, 2025 at 8:23 PM
I love hearing feedback like this about the new RAG course ❤️
To learn more about the course go here: elite-ai-assisted-coding.dev/p/learn-rag...
To learn more about the course go here: elite-ai-assisted-coding.dev/p/learn-rag...
80% of users won’t look past the top 5 results. If those results are bad, they leave.
I just launched a RAG course, here’s a post on what you'll learn and why the concepts matter in modern AI systems
elite-ai-assisted-coding.dev/p/learn-rag...
I just launched a RAG course, here’s a post on what you'll learn and why the concepts matter in modern AI systems
elite-ai-assisted-coding.dev/p/learn-rag...
Learn RAG: Build Search That Works
Search isn’t magic. It’s engineering. 80% of users won’t look past the top 5 results. If those results are bad, they leave.
elite-ai-assisted-coding.dev
October 9, 2025 at 4:14 PM
80% of users won’t look past the top 5 results. If those results are bad, they leave.
I just launched a RAG course, here’s a post on what you'll learn and why the concepts matter in modern AI systems
elite-ai-assisted-coding.dev/p/learn-rag...
I just launched a RAG course, here’s a post on what you'll learn and why the concepts matter in modern AI systems
elite-ai-assisted-coding.dev/p/learn-rag...
Building in prior version rollback for AI agents is a pain (code and db).
Neon’s new snapshots API solves this.
Save a complete snapshot of your db, schema, and data with each agent turn. It's efficiently done with copy-on-write vs full copies too.
elite-ai-assisted-coding.dev/p/build-che...
Neon’s new snapshots API solves this.
Save a complete snapshot of your db, schema, and data with each agent turn. It's efficiently done with copy-on-write vs full copies too.
elite-ai-assisted-coding.dev/p/build-che...
Build Checkpoints for Your Agent with Neon Snapshots
If you build apps with an AI agent, you probably want a way to go back in time to prior agent turns.
elite-ai-assisted-coding.dev
October 3, 2025 at 3:58 PM
Building in prior version rollback for AI agents is a pain (code and db).
Neon’s new snapshots API solves this.
Save a complete snapshot of your db, schema, and data with each agent turn. It's efficiently done with copy-on-write vs full copies too.
elite-ai-assisted-coding.dev/p/build-che...
Neon’s new snapshots API solves this.
Save a complete snapshot of your db, schema, and data with each agent turn. It's efficiently done with copy-on-write vs full copies too.
elite-ai-assisted-coding.dev/p/build-che...
I used AI to turn a hacky prototype into a full app in one day and understood every line of code
This workflow is for using AI to map out a plan, make architectural decisions, implement features step-by-step. And I code and refactor where it matters
elite-ai-assisted-coding.dev/p/using-ai-...
This workflow is for using AI to map out a plan, make architectural decisions, implement features step-by-step. And I code and refactor where it matters
elite-ai-assisted-coding.dev/p/using-ai-...
Using AI For Real Projects and Decisions
I had an idea, and a hacky prototype I wanted to build.
elite-ai-assisted-coding.dev
October 1, 2025 at 4:01 PM
I used AI to turn a hacky prototype into a full app in one day and understood every line of code
This workflow is for using AI to map out a plan, make architectural decisions, implement features step-by-step. And I code and refactor where it matters
elite-ai-assisted-coding.dev/p/using-ai-...
This workflow is for using AI to map out a plan, make architectural decisions, implement features step-by-step. And I code and refactor where it matters
elite-ai-assisted-coding.dev/p/using-ai-...
Code with AI smarter, not harder ⚡
This free 100+ page book is packed with pro-tips, step-by-step guides, and solutions to accelerate your workflow.
Get your copy 👇
maven.com/p/aa2670/ti...
This free 100+ page book is packed with pro-tips, step-by-step guides, and solutions to accelerate your workflow.
Get your copy 👇
maven.com/p/aa2670/ti...
Tips, Guides, and Answered Questions by Eleanor Berger and Isaac Flath on Maven
Over 100 pages of answered questions, guides, and tips and tricks from Isaac
Flath and Eleanor Berger!
maven.com
September 30, 2025 at 4:01 PM
Code with AI smarter, not harder ⚡
This free 100+ page book is packed with pro-tips, step-by-step guides, and solutions to accelerate your workflow.
Get your copy 👇
maven.com/p/aa2670/ti...
This free 100+ page book is packed with pro-tips, step-by-step guides, and solutions to accelerate your workflow.
Get your copy 👇
maven.com/p/aa2670/ti...
AI suggested a dependency I didn't understand. Instead of accepting or rejecting, I asked why
The debate and research taught me more about my async stack than a tutorial ever could
AI suggestions are opportunities to learn
Details + more examples 👇
elite-ai-assisted-coding.dev/p/i-argue-w...
The debate and research taught me more about my async stack than a tutorial ever could
AI suggestions are opportunities to learn
Details + more examples 👇
elite-ai-assisted-coding.dev/p/i-argue-w...
I Argue With My AI Pair Programmer, and It Makes My Code Better
Sometimes AI teaches me something and sometimes AI is a better rubber duck. Both are useful
elite-ai-assisted-coding.dev
September 29, 2025 at 5:09 PM
AI suggested a dependency I didn't understand. Instead of accepting or rejecting, I asked why
The debate and research taught me more about my async stack than a tutorial ever could
AI suggestions are opportunities to learn
Details + more examples 👇
elite-ai-assisted-coding.dev/p/i-argue-w...
The debate and research taught me more about my async stack than a tutorial ever could
AI suggestions are opportunities to learn
Details + more examples 👇
elite-ai-assisted-coding.dev/p/i-argue-w...
Join us and more than 1,600 others next week as we dive into building a coding agent from scratch!
maven.com/p/b9a2df/bu...
maven.com/p/b9a2df/bu...
Build Your Own AI Coding Agent from Scratch
In the spirit of Feynman’s “What I cannot create, I do not understand,” this optional outtake from our course is a live mini-build: we’ll code a simple AI coding agent together to form a real mental model and better understand the tools we use. We’ll practice context management, clear specs, small steps, and interactive + structured flows. Join to watch — or hack along.
maven.com
September 29, 2025 at 4:02 PM
Join us and more than 1,600 others next week as we dive into building a coding agent from scratch!
maven.com/p/b9a2df/bu...
maven.com/p/b9a2df/bu...
What happens when a whole engineering team learns to use AI together?
They build a common language & framework for prompts, context, tools, and automations. This mean effective collaboration, faster problem-solving, and code that is consistent across your team.
They build a common language & framework for prompts, context, tools, and automations. This mean effective collaboration, faster problem-solving, and code that is consistent across your team.
September 27, 2025 at 7:57 PM
What happens when a whole engineering team learns to use AI together?
They build a common language & framework for prompts, context, tools, and automations. This mean effective collaboration, faster problem-solving, and code that is consistent across your team.
They build a common language & framework for prompts, context, tools, and automations. This mean effective collaboration, faster problem-solving, and code that is consistent across your team.
GitHub's new Copilot CLI gives:
✅ Terminal agent interface like claude code, opencode, gemini CLI
✅ A more transparent and predictable "per-request" billing model
See how using it to create and use an MCP server went👇
elite-ai-assisted-coding.dev/p/github-co...
✅ Terminal agent interface like claude code, opencode, gemini CLI
✅ A more transparent and predictable "per-request" billing model
See how using it to create and use an MCP server went👇
elite-ai-assisted-coding.dev/p/github-co...
GitHub Copilot CLI Tool Review
Your Copilot subscription now comes with a dedicated terminal-based agent
elite-ai-assisted-coding.dev
September 26, 2025 at 4:03 PM
GitHub's new Copilot CLI gives:
✅ Terminal agent interface like claude code, opencode, gemini CLI
✅ A more transparent and predictable "per-request" billing model
See how using it to create and use an MCP server went👇
elite-ai-assisted-coding.dev/p/github-co...
✅ Terminal agent interface like claude code, opencode, gemini CLI
✅ A more transparent and predictable "per-request" billing model
See how using it to create and use an MCP server went👇
elite-ai-assisted-coding.dev/p/github-co...