
Git Maxd
@GitMaxd
Followers
1K
Following
25K
Media
473
Statuses
3K
Linux Dev | AI Rev Early Adopter | Love Next | LangChain 🦜 | Gen AI\UX
Phoenix, AZ
Joined July 2023
I just shipped o3-mcp as an OSS 🚀. The power of the @OpenAI o3 Model family in your favorite IDE or MCP Client as a collaborative Pair Agent!. o3-pro Model included 🔥. You can now initiate a prompt or chat dialog between your model of choice and the o3 Models! . npx o3-mcp 👇
7
4
48
Really great live stream just started with @pvncher & @RayFernando1337 . Talking about @RepoPrompt - one of the best opportunities to learn straight from the creator 🚀. Wondering about RepoPrompt? This is it!.
0
4
9
One of the better explanations of how Cursor gathers context in order to more efficiently use tokens. This may still be the best way for “Vibe” coding given that it’s relatively effective at selecting context. Less expensive for the provider at the consumer expense.
@aidan_mclau Cursor uses embedding models in place of search + grep and doesn’t let their agent read the full file when looking around for context. Context is everything, and cursor agent is starved for it.
1
0
3
@LangChainAI @tavilyai Or, you can just run `npm create langgraph@latest` and select the template of your choice ❤️.
0
0
0
How to create a @LangChainAI LangGraph JS ReAct Agent. 1. git clone 2. Add AI key to .env.3. yarn install.4. npx @langchain/langgraph-cli dev. You'll have a @tavilyai search enabled configurable LangGraph Agent Running in LangGraph Studio. Iterate & Vibe
2
1
9
Ran across a great talk by @dexhorthy giving his framework agnostic explanation of “Context Engineering” 12 Factor Agents. I’ve been focusing a lot on Context engineering for a Context Creation system I’ve been working on and this talk has a lot of agentic context goodness 🔥
"everything that makes agents good is context engineering" . excited to release @dexhorthy's talk at @aiDotEngineer, coiner of Context Engineering which has captured the zeitgeist of some of the most important problems in AI Engineering today!
1
2
8
Pretty big deal right here. Use @RepoPrompt as an MCP tool to edit your project files. Work in Claude Desktop and use the brilliant Context Engineering in Repo Prompt to provide both Single and Multi Edit modes. It’s a very interesting development 🔥
.@RepoPrompt 1.3.12 is live!.MCP Server Wave 2 → File Operations. - Create, delete, move, and edit files. - Apply multiple search-and-replace operations in a single edit. - Revamped search tool to allow restricted search area.- Consolidated tool definitions!. Give it a try!.
1
2
13
. @RLanceMartin (Lance from @LangChainAI) just published an excellent blog and YT on Context Engineering. It’s timely information on how Agents are being used to create context (ie Memories) as well as the dangers that may lie ahead. Bookmark it 🔖.
Context Engineering. "Context engineering" is a key part of agent building. We review a few popular patterns for context engineering + explain how to use them w/ LangGraph. Blog:. Video:.
0
2
10
Excited to see this new lesson in @LangChainAI Academy. Learn about LangGraph Ambient Agents in this free 27 lesson, 2.5hr video course presented by @RLanceMartin . Includes Agent/Human In The Loop example 🔥
Our latest LangChain Academy course – Building Ambient Agents with LangGraph – is now available!. Most agents today handle one request at a time through chat interfaces. But as models have improved, agents can now run in the background – and take on long-running, complex tasks.
1
3
18
“Context engineering is the delicate art and science of filling the context window with just the right information for the next step”. It is artistic.
+1 for "context engineering" over "prompt engineering". People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window.
0
1
3
This is excellent prompting advice by @BraceSproul . Prompting the model to do something is more than just asking. There’s a lot of planning here. This flows straight into the larger Content Engineering concepts we’ve been reading about.
Today I'm going to talk about conversation history management and summarization. The example below comes from the same `planner` sub-agent discussed yesterday. This is one of the many sub-agents which make up a coding agent we'll be open sourcing soon!. The `planner` agent is.
0
0
5
If you love learning smart things, this is for you ❤️🔥. There’s a lot going on here. 1. @firecrawl_dev Batch API.2. New @upstash “Search” service.3. OpenAI Compatible API example. These Firecrawl demos are a goldmine of context, examples, and HowTo’s.
Announcing Firestarter - an open-source Chatbot building platform 🔥. Crawl websites, train chatbots on multiple sources, and automatically generate endpoints for every bot. Try out the perfect chatbot starter kit today 👇
3
5
23
You are 100% responsible for your own memories. You can’t make this up. Do memory systems belong in an IDE designed cut context?.
just gonna say it, the current @cursor_ai "memories" is terrible and puts the worst stuff possible and ai doesn't even use them. I know its in beta but still this is pathetic. .
1
0
2
The cost efficiencies you get with Claude Code and the @AnthropicAI Max Sub may be the best deal rn. 5/24 - 6/12 . I was only Opus limited, late in the day, one time. If you use it, it’s amazing for $200/mo.
0
0
4
RT @GitMaxd: @RayFernando1337 When you start to understand what Claude Code rly is, it changes things. I use Claude Code as an MCP server….
0
2
0
Even using the best web scraper to ever exist, @firecrawl_dev 🔥, I still like to trim the gristle out of my context. Context should be healthy, not bloated with extraneous data. Context Clean Up, Isle 3.
0
0
7
RT @GitMaxd: @RayFernando1337 @firecrawl_dev o3 in Claude Desktop - pretty great. Works exactly the same in Claude Code .
0
2
0