[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.] #  @svpino Santiago Santiago posts on X about ai, $googl, if you, prompt engineering the most. They currently have XXXXXXX followers and XX posts still getting attention that total XXXXXXX engagements in the last XX hours. ### Engagements: XXXXXXX [#](/creator/twitter::18737039/interactions)  - X Week XXXXXXXXX -XXXX% - X Month XXXXXXXXX +22% - X Months XXXXXXXXXX -XX% - X Year XXXXXXXXXX -XX% ### Mentions: XX [#](/creator/twitter::18737039/posts_active)  - X Week XX -XX% - X Month XXX +15% - X Months XXX +21% - X Year XXXXX -XXXX% ### Followers: XXXXXXX [#](/creator/twitter::18737039/followers)  - X Week XXXXXXX +0.11% - X Month XXXXXXX +0.49% - X Months XXXXXXX +4.20% - X Year XXXXXXX +12% ### CreatorRank: XXXXXX [#](/creator/twitter::18737039/influencer_rank)  ### Social Influence **Social category influence** [travel destinations](/list/travel-destinations) #1363 [technology brands](/list/technology-brands) #5611 [stocks](/list/stocks) #1009 [finance](/list/finance) XXXX% [social networks](/list/social-networks) XXXX% **Social topic influence** [ai](/topic/ai) #1867, [$googl](/topic/$googl) #130, [if you](/topic/if-you) 11.48%, [prompt engineering](/topic/prompt-engineering) #6, [agentic](/topic/agentic) #4, [context engineering](/topic/context-engineering) #4, [the first](/topic/the-first) #3312, [science](/topic/science) #302, [focusing](/topic/focusing) 3.28%, [insurance](/topic/insurance) #632 **Top accounts mentioned or mentioned by** [@rifadm816](/creator/undefined) [@1concordia](/creator/undefined) [@rutibex](/creator/undefined) [@seventhwhiskey](/creator/undefined) [@37signals](/creator/undefined) [@ghostov_aldavis](/creator/undefined) [@zenul_abidin](/creator/undefined) [@rtheoryxyz](/creator/undefined) [@maximilianog](/creator/undefined) [@anipiano3](/creator/undefined) [@danbestor](/creator/undefined) [@club_susan](/creator/undefined) [@cooperchurch427](/creator/undefined) [@workos](/creator/undefined) [@nextjs](/creator/undefined) [@karpathy](/creator/undefined) [@secondmeai](/creator/undefined) [@ocrarena](/creator/undefined) [@flowith](/creator/undefined) [@trakpalnet](/creator/undefined) **Top assets mentioned** [Alphabet Inc Class A (GOOGL)](/topic/$googl) ### Top Social Posts Top posts by engagements in the last XX hours "Google: "Our productivity has increased XX% thanks to AI." Random Internet User: "Skill Issue We are 100x faster" This is a tough one: I'm not sure who to believe here" [X Link](https://x.com/svpino/status/1974809297461911749) 2025-10-05T12:10Z 431K followers, 35K engagements "Here is an agent that literally runs on your computer to get stuff done. Not in the cloud. Not in a sandbox. Check out (Simular 1). It can open apps browse the web find and modify files and do things for you while you buy groceries at the store. I know this sounds scary but here is the good news: You can train this agent to do only what you need it to do. Basically you walk the agent through a task so it learns how to do it. From there it can take over and do it for you. This can pretty much automate anything you do on your computer. Pretty crazy where this is going (Their video is pretty" [X Link](https://x.com/svpino/status/1995888276566409273) 2025-12-02T16:10Z 431K followers, 33.5K engagements "MCP now supports authentication using OAuth2. This is huge I want to share a complete example (code) so you can learn how to use it. Here is the thing: You don't want to build agentic applications that access resources without proper guardrails. With authentication you can: Call third-party APIs on the user's behalf Perform actions that should be traceable Enforce different permissions for different roles Here is an MCP-based web shop demo built with @WorkOS AuthKit and @nextjs. With this example you can do the following: Connect to an MCP server Authenticate using AuthKit Order a T-shirt" [X Link](https://x.com/svpino/status/1996251153865244766) 2025-12-03T16:12Z 431K followers, 34K engagements "Context engineering is the new prompt engineering. Prompts aren't the bottleneck anymore. Agents fail because their context is noisy overloaded or irrelevant. Bigger context windows don't fix bad context. This is one of the reasons everyone who's claimed that RAG is dead was mistaken. Bad lacking or poorly constructed context = bad agent decisions. The perfect prompt + bad context = bad agent decisions. The best agent developers today aren't the best at writing prompts. They are attention architects. That's where real reliability comes from" [X Link](https://x.com/svpino/status/1996567683543036039) 2025-12-04T13:10Z 431K followers, 25.5K engagements "This is an excellent book. @karpathy said: "Context engineering is the delicate art and science of filling the context window with just the right information for the next step." This book will help you stop thinking about "prompt engineering" and start focusing on "context engineering" instead. This is the book I'd recommend to anyone looking to become an "AI Solutions Architect". Amazon link:" [X Link](https://x.com/svpino/status/1993047496558522814) 2025-11-24T20:02Z 431K followers, 128.6K engagements "We are about to cross the uncanny valley. It's getting really hard to tell whether videos are real. This year you can have your kids talk to an AI-generated Santa and blow their minds. x This is not a prerecorded call and there's nothing scripted. It just feels real. It can pick up on your facial expressions and respond to your tone of voice. It can also adapt to different topics and reacts very well when the conversation goes off the rails. Here is the link: We aren't ready for what's coming" [X Link](https://x.com/svpino/status/1993772277436174359) 2025-11-26T20:02Z 431K followers, 10.1K engagements "This is literally the easiest way to build an AI agent. Zero code. You only need to run a couple of commands. (I've said this before but Google ADK is my favorite way to build code-first agents.) Here is what you need to do: X. Open your terminal and create an empty folder X. Run uv init inside the folder (if you don't have uv installed you are really missing out) X. Install the google-adk library using the following command: uv add google-adk. X. At this point you can use the ask to create your agent with a single command: uv run adk create --type=config my_agent. This will ask you a couple" [X Link](https://x.com/svpino/status/1996209124573651077) 2025-12-03T13:25Z 431K followers, 137.3K engagements "This is one of the craziest concepts I've seen so far: Second Me is a platform that creates an AI identity based on you: It takes your photos It takes your voice It takes your notes And it creates a second you (a virtual copy). It's an AI-powered identity that sounds and acts like you. I'm actually not surprised at all. I saw this coming a mile away. Back in the day the Internet was all about wandering into random forums to meet people who thought like you. Now it's all about algorithms. That's where Second Me wants to play: You can take your generated AI identity and let it go out and talk" [X Link](https://x.com/svpino/status/1996673518973734932) 2025-12-04T20:10Z 431K followers, 297K engagements "People saying you can vibe-code production-ready applications are lying to you. Real-world production applications inevitably lead to complex code. Lots of code you'll have to write and maintain. Absolutely zero chance you'll get away with writing prompts" [X Link](https://x.com/svpino/status/1993672597792518177) 2025-11-26T13:26Z 431K followers, 233.2K engagements "Google is cooking on all cylinders It's just not Gemini 3which is awesome But they are putting a lot of work into the ADK (Agent Development Kit). Here is just one example of how they are focusing on development experience: You build your agent. You XXX% focus on your code and make it work locally. When you are done you can use Google's agent-starter-pack to add everything you need to make it production-ready. One command and you'll get infrastructure CI/CD observability and security: $ uvx agent-starter-pack enhance Pretty magic stuff" [X Link](https://x.com/svpino/status/1998389939860607121) 2025-12-09T13:51Z 431K followers, 65.4K engagements "I still think that prompt engineering is not a career. I still think companies will not start hiring prompt engineers anytime soon. I still think about prompt engineering as a skill not a role" [X Link](https://x.com/svpino/status/1971559145431257323) 2025-09-26T12:55Z 431.1K followers, 15.8K engagements "Heard somebody call a dude talking about vibe coding you are a glorified prompt-typer and I cant stop laughing" [X Link](https://x.com/svpino/status/1985853290316284339) 2025-11-04T23:34Z 431K followers, 8652 engagements "Stop replying to social media posts using AI. It's a pretty stupid strategy that makes everything worse for you and everyone else" [X Link](https://x.com/svpino/status/1995543743886491689) 2025-12-01T17:21Z 431.1K followers, 28.9K engagements "My health insurance in the US. Family of X. $15000 deductible. $20000 out-of-pocket. You can't call this health insurance. This is just to protect against a catastrophe. We are the dumbest country in the world that can't figure out how to fix this broken system" [X Link](https://x.com/svpino/status/1995614596745625875) 2025-12-01T22:02Z 431K followers, 661.9K engagements "Prompt engineering is dead. Context engineering is the new king. I've talked to many people building agents and most are duct-taped together using a bunch of services: One to store messages One to manage memory Artifacts in S3 A database acting as a journal This is a complex mess. Agents don't learn from their experience. They can't observe their own behavior and do something with it because it's all scattered everywhere. Fortunately we are trying to solve this: Check out Acontext (GitHub Repository link below). This is the first platform I've seen that's centered around context: STORE what" [X Link](https://x.com/svpino/status/1998785305038712841) 2025-12-10T16:02Z 431K followers, 22.4K engagements "Reviewing AI-generated code is the new bottleneck. Writing code was slow before. Now in a few minutes you can generate virtually unlimited lines of code using AI. But developers are now spending XX% of their time checking that code. That's the new bottleneck. I've been using Cline and they have a feature that actually helps with this: You can click a button and have Cline generate an inline explanation for every single line of AI-generated code. Take a look at the attached video. What it does: X. You can read the code and the explanation together X. You don't lose context while reviewing code" [X Link](https://x.com/svpino/status/1996610800040157642) 2025-12-04T16:01Z 431K followers, 63.4K engagements "What a stupid timeline we are living in. Beyond parody" [X Link](https://x.com/svpino/status/1997027456843231730) 2025-12-05T19:37Z 431K followers, 47.5K engagements "You don't need another tutorial. You don't learn football by watching videos or reading a rule book. You learn by running and kicking the ball. Stop bookmarking webpages and downloading books. Writing code and building a project is the only way to learn" [X Link](https://x.com/svpino/status/1997656095301767359) 2025-12-07T13:15Z 431.1K followers, 19.2K engagements "My mom is obsessed with AI. She has taken her entire photo catalog and generated 1000 variations of all of the family photos using Dreamina + Nano Banana Pro. Probably one of the best Christmas gifts you can give someone is to show them how to do this" [X Link](https://x.com/svpino/status/1998484065348182253) 2025-12-09T20:05Z 431K followers, 12.1K engagements "If you can't see what an agent does you can't improve it you can't debug it and you can't trust it. It's crazy how many teams are building agents with no way to understand what they're doing. Literally ZERO observability. This is probably one of the first questions I ask every new team I meet: Can you show me the traces of a few executions of your agents Nada. Zero. Nilch. Large language models make bad decisions all the time. Agents fail and you won't realize it until somebody complains. At a minimum every agent you build should produce traces showing the full request flow latency analysis" [X Link](https://x.com/svpino/status/1998754848351859058) 2025-12-10T14:01Z 431K followers, 14.4K engagements "it sure looks like google is winning" [X Link](https://x.com/svpino/status/1998816787190685716) 2025-12-10T18:07Z 431K followers, 22.1K engagements "This is insane power for software developers: An agentic principal engineer that you can install as an MCP server. It can index any codebase and surface insights you didn't even know about. This is likely the most useful MCP server you'll start using. It's called @Qodo Aware and it's the first production-ready deep research agent built for navigating and understanding enterprise-scale codebases. Find the GitHub repository in a comment below. The beauty of this is that you don't have to switch tools to take advantage of it: you can keep using Cursor Visual Studio Code Claude Code or whatever" [X Link](https://x.com/svpino/status/1965855311451009254) 2025-09-10T19:10Z 431K followers, 78.6K engagements "I paid $150/mo for a mobile developer. That was around X years ago. Today you can describe the app you want to create in plain English and have AI build it for you. Check out the video I recorded using OnSpace AI. It took me XX minutes to build a mobile application that scans receipts and uses OCR to read all of the information. A few of the highlights: You can create what you want by just describing it. You can plug in cutting-edge AI models instantly. No API keys no setup. You can stitch together reasoning workflows data and UI without writing a single line of code. Whatever you build runs" [X Link](https://x.com/svpino/status/1994074373477818494) 2025-11-27T16:02Z 431K followers, 45.1K engagements "There is a massive opportunity right now for anyone who can walk into a company and automate the grunt work they do every day. It doesn't matter what programming language you use or libraries you like or whether you know what micro-services do or if you like Claude Code or Cursor. No one cares about any of that. If you can identify a problem and convince them you can solve it you'll make a killing" [X Link](https://x.com/svpino/status/1994402155906859223) 2025-11-28T13:45Z 431K followers, 49.8K engagements "Running LLMs at scale is just too expensive. This is a common concern for most companies I talk to. We need faster and cheaper models. Here is a FREE session where the Nebius team will break down LLM distillation and give you an end-to-end practical guide on how to use it. You'll learn: How to train smaller models that stay fast and accurate How to cut inference costs without painful trade-offs How to deploy distilled models in production with SLAs and zero-retention guarantees For those of you who are hands-on you'll get a full demo that will walk you through the entire workflow. If you're" [X Link](https://x.com/svpino/status/1995584960238064089) 2025-12-01T20:05Z 431K followers, 15.6K engagements "I used to charge $20000 per week to come and help your team fix your software architecture and revive your failed projects. A big part of that was finding and fixing bugs. It's mind-blowing to realize you can now do a lot of the same work for $XXX. Here is a tip: Check out Max at Max is an "AI software engineer" that does something I haven't seen done yet: X. It helps you build your application X. The agent then loads and uses the app autonomously to find and fix bugs 👀 I think this is new and it's got huge potential Max loads your app clicks around types and uses it to test it. This is" [X Link](https://x.com/svpino/status/1996312316271546573) 2025-12-03T20:15Z 431.1K followers, 31.8K engagements "To me open source has always been about "opening the source code." That's literally been my understanding since forever. Well it turns out that somebody decided that "open source" doesn't mean that and instead they have a purity test to define whether you can call something "open" or not. If you subscribe to this redefinition of the term I'm happy for you. But you are absolutely nobody to tell me what open source is and how and when I can use the term" [X Link](https://x.com/svpino/status/1996992437256855561) 2025-12-05T17:17Z 431K followers, 13.5K engagements "Here is what I like about Google ADK to build code-first agents: X. Open-source X. Backed by Google X. Model-agnostic - It works with Gemini OpenAI models Claude Llama or whatever other model you want to use. X. Multi-agent by design - You can start with a simple agent and evolve your system to as many specialized agents as you need. X. Code-first - This is all Python and you have complete control over everything (instead of depending on a tool that hides the code behind a drag & drop interface.) X. You get a CLI and a visual web UI for debugging and evaluation. X. Enterprise-ready from day 1" [X Link](https://x.com/svpino/status/1996933834429079812) 2025-12-05T13:25Z 431K followers, 16.9K engagements "We have X new benchmark datasets to evaluate LLMs' structured-output capabilities. (Because existing benchmarks are full or errors) X. Data Table Analysis - Evaluates how well an LLM extracts key metadata about each table. X. Insurance Claims Extraction - Evaluates how well an LLM extracts information about each claim. X. Financial Entities Extraction - Evaluates how well an LLM extracts structured financial and contextual entities mentioned in a document. X. PII Extraction - Evaluates how well an LLM extracts different categories of PII from the document. This is awesome" [X Link](https://x.com/svpino/status/1997002700718784870) 2025-12-05T17:58Z 431K followers, 11.3K engagements "This is still the way I recommend most people start with machine learning: X. Start with Python X. Learn to use Google Colab X. Take a Pandas tutorial X. Then a Seaborn tutorial X. Learn how to use Decision Trees X. Finish Kaggle's "Intro to Machine Learning" X. Solve the Titanic challenge" [X Link](https://x.com/svpino/status/1979929321667252466) 2025-10-19T15:15Z 431K followers, 461.6K engagements "Best GPU prices you'll find out there: H100 GPU for $2.00/hr H200 GPU for $2.30/hr Even better: You can literally move all of your data to Nebius *for free* including a free PoC with support and get up to X months for free Nebius is collaborating with me on this post and their deal is just amazing This is insanely cheap and you should check it out if you are looking to fine-tune models run experiments or do some production testing. Check this out:" [X Link](https://x.com/svpino/status/1970143734408101970) 2025-09-22T15:10Z 431K followers, 59.2K engagements "Web scraping is a critical skill for software developers At a time when every large company is training large language models learning how to scrape content from the web is a top skill to learn. Web scraping Collecting data at scale and storing it in a structured format like CSV or JSON. Here is how the process works at a high level: You send a request to the URL you want to scrape The server responds with the HTML of the page Your code parses the HTML and collects the data Rinse and repeat for every URL from which you want to scrape data. Here is a free and open-source tool you can use:" [X Link](https://x.com/svpino/status/1973043933006754169) 2025-09-30T15:15Z 431.1K followers, 154.1K engagements "Ive been building software for XX years more than XX professionally. I understand architecture principles and best practices. No you cant vibe-code Salesforce SAP or any other complex enterprise app" [X Link](https://x.com/svpino/status/1993762257080033721) 2025-11-26T19:22Z 431K followers, 372.8K engagements "I don't believe you can vibe-code complex applications. Put another way your ability to vibe-code software is inversely proportional to its complexity. Vibe-coding is like going into the "Minotaur Labyrinth" without Theseus' thread: you'll invariably get lost. More accurately you'll be stuck in an endless loop going in circles with the model: Things that used to work stop working Your code starts bloating Solutions become over-engineered And all of this will happen while working with popular tech where models have plenty of training. Try to step outside the beaten path (use anything a bit" [X Link](https://x.com/svpino/status/1994031036720926881) 2025-11-27T13:10Z 431K followers, 77.7K engagements "Imagine getting mad because @37signals opened the source code of their newest product but kept the right to monetize it. I will never understand people like this. If they kept the source code closed (like literally every single saas builder does) nobody would have complained. But they dared to open-source that code so that's a problem somehow @37signals is a gift to our industry. We should have more companies doing as they do not fewer" [X Link](https://x.com/svpino/status/1996634183465152856) 2025-12-04T17:34Z 431.1K followers, 65.5K engagements "Kiro keeps getting better For those of you who don't know Kiro is the AI-powered coding assistant from AWS. They just released "Kiro Powers" and here is how it works: Instead of dumping every MCP tool into context and hoping your agent figures it out this feature helps Kiro load exactly what you need exactly when you need it. The whole idea is to prevent context rot and prevent having to create rules and custom instructions by hand. You now have a POWER .md file that acts as the onboarding manual the agent reads. It tells the agent what tools to use when to use them and how to use them" [X Link](https://x.com/svpino/status/1996975378729148629) 2025-12-05T16:10Z 431K followers, 12.4K engagements "I think it's time to start talking about AgentOps. DevOps MLOps AgentOps If you want autonomous agents that work and scale we need to start formalizing the discipline that supports them. Some of the things *everyone* has to worry about: Agent evaluations (using LLM-as-a-judge or any other technique) Metrics-driven releases (ship new versions when they beat the threshold) A/B testing agents canary releases shadow deployments Enabling observability and tracing Implementing and integrating human feedback loops You can't build a reliable agent without worrying about everything above" [X Link](https://x.com/svpino/status/1997305280271659227) 2025-12-06T14:01Z 431K followers, 27.7K engagements "Stripe wanted to buy this company but the founder declined their $XXX billion offer. Today the global financial system is about to change because of what Airwallex has been doing. Now they are a $1B ARR global fintech and they're building what could be the first execution layer for autonomous finance. These AI agents can now do what used to take hours: Pull receipts from emails Slack WhatsApp Match them to transactions Categorize automatically Catch policy exceptions before they slip through The underlying infrastructure they are using is pretty impressive: XX% of payments run on local rails" [X Link](https://x.com/svpino/status/1998044050348540348) 2025-12-08T14:56Z 431K followers, 40.6K engagements "This story is incredible: Back in 2005 a guy switched to Computer Science because he wanted to work at Google. After a ton of work he finally got an internship at the company. When he left he founded his first company. Google backed it. He then built something far more ambitious: Emergent an end-to-end agentic vibe-coding platform. And he just closed funding from Google's AI Futures Fund As of today Emergent has: 2.5M+ people using it $25M ARR (they got there in X months) 3M+ apps shipped You can use Emergent to create fully functioning products with a prompt: X. The front-end application 2." [X Link](https://x.com/svpino/status/1998847717234319780) 2025-12-10T20:10Z 431K followers, 137.4K engagements
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
@svpino SantiagoSantiago posts on X about ai, $googl, if you, prompt engineering the most. They currently have XXXXXXX followers and XX posts still getting attention that total XXXXXXX engagements in the last XX hours.
Social category influence travel destinations #1363 technology brands #5611 stocks #1009 finance XXXX% social networks XXXX%
Social topic influence ai #1867, $googl #130, if you 11.48%, prompt engineering #6, agentic #4, context engineering #4, the first #3312, science #302, focusing 3.28%, insurance #632
Top accounts mentioned or mentioned by @rifadm816 @1concordia @rutibex @seventhwhiskey @37signals @ghostov_aldavis @zenul_abidin @rtheoryxyz @maximilianog @anipiano3 @danbestor @club_susan @cooperchurch427 @workos @nextjs @karpathy @secondmeai @ocrarena @flowith @trakpalnet
Top assets mentioned Alphabet Inc Class A (GOOGL)
Top posts by engagements in the last XX hours
"Google: "Our productivity has increased XX% thanks to AI." Random Internet User: "Skill Issue We are 100x faster" This is a tough one: I'm not sure who to believe here"
X Link 2025-10-05T12:10Z 431K followers, 35K engagements
"Here is an agent that literally runs on your computer to get stuff done. Not in the cloud. Not in a sandbox. Check out (Simular 1). It can open apps browse the web find and modify files and do things for you while you buy groceries at the store. I know this sounds scary but here is the good news: You can train this agent to do only what you need it to do. Basically you walk the agent through a task so it learns how to do it. From there it can take over and do it for you. This can pretty much automate anything you do on your computer. Pretty crazy where this is going (Their video is pretty"
X Link 2025-12-02T16:10Z 431K followers, 33.5K engagements
"MCP now supports authentication using OAuth2. This is huge I want to share a complete example (code) so you can learn how to use it. Here is the thing: You don't want to build agentic applications that access resources without proper guardrails. With authentication you can: Call third-party APIs on the user's behalf Perform actions that should be traceable Enforce different permissions for different roles Here is an MCP-based web shop demo built with @WorkOS AuthKit and @nextjs. With this example you can do the following: Connect to an MCP server Authenticate using AuthKit Order a T-shirt"
X Link 2025-12-03T16:12Z 431K followers, 34K engagements
"Context engineering is the new prompt engineering. Prompts aren't the bottleneck anymore. Agents fail because their context is noisy overloaded or irrelevant. Bigger context windows don't fix bad context. This is one of the reasons everyone who's claimed that RAG is dead was mistaken. Bad lacking or poorly constructed context = bad agent decisions. The perfect prompt + bad context = bad agent decisions. The best agent developers today aren't the best at writing prompts. They are attention architects. That's where real reliability comes from"
X Link 2025-12-04T13:10Z 431K followers, 25.5K engagements
"This is an excellent book. @karpathy said: "Context engineering is the delicate art and science of filling the context window with just the right information for the next step." This book will help you stop thinking about "prompt engineering" and start focusing on "context engineering" instead. This is the book I'd recommend to anyone looking to become an "AI Solutions Architect". Amazon link:"
X Link 2025-11-24T20:02Z 431K followers, 128.6K engagements
"We are about to cross the uncanny valley. It's getting really hard to tell whether videos are real. This year you can have your kids talk to an AI-generated Santa and blow their minds. x This is not a prerecorded call and there's nothing scripted. It just feels real. It can pick up on your facial expressions and respond to your tone of voice. It can also adapt to different topics and reacts very well when the conversation goes off the rails. Here is the link: We aren't ready for what's coming"
X Link 2025-11-26T20:02Z 431K followers, 10.1K engagements
"This is literally the easiest way to build an AI agent. Zero code. You only need to run a couple of commands. (I've said this before but Google ADK is my favorite way to build code-first agents.) Here is what you need to do: X. Open your terminal and create an empty folder X. Run uv init inside the folder (if you don't have uv installed you are really missing out) X. Install the google-adk library using the following command: uv add google-adk. X. At this point you can use the ask to create your agent with a single command: uv run adk create --type=config my_agent. This will ask you a couple"
X Link 2025-12-03T13:25Z 431K followers, 137.3K engagements
"This is one of the craziest concepts I've seen so far: Second Me is a platform that creates an AI identity based on you: It takes your photos It takes your voice It takes your notes And it creates a second you (a virtual copy). It's an AI-powered identity that sounds and acts like you. I'm actually not surprised at all. I saw this coming a mile away. Back in the day the Internet was all about wandering into random forums to meet people who thought like you. Now it's all about algorithms. That's where Second Me wants to play: You can take your generated AI identity and let it go out and talk"
X Link 2025-12-04T20:10Z 431K followers, 297K engagements
"People saying you can vibe-code production-ready applications are lying to you. Real-world production applications inevitably lead to complex code. Lots of code you'll have to write and maintain. Absolutely zero chance you'll get away with writing prompts"
X Link 2025-11-26T13:26Z 431K followers, 233.2K engagements
"Google is cooking on all cylinders It's just not Gemini 3which is awesome But they are putting a lot of work into the ADK (Agent Development Kit). Here is just one example of how they are focusing on development experience: You build your agent. You XXX% focus on your code and make it work locally. When you are done you can use Google's agent-starter-pack to add everything you need to make it production-ready. One command and you'll get infrastructure CI/CD observability and security: $ uvx agent-starter-pack enhance Pretty magic stuff"
X Link 2025-12-09T13:51Z 431K followers, 65.4K engagements
"I still think that prompt engineering is not a career. I still think companies will not start hiring prompt engineers anytime soon. I still think about prompt engineering as a skill not a role"
X Link 2025-09-26T12:55Z 431.1K followers, 15.8K engagements
"Heard somebody call a dude talking about vibe coding you are a glorified prompt-typer and I cant stop laughing"
X Link 2025-11-04T23:34Z 431K followers, 8652 engagements
"Stop replying to social media posts using AI. It's a pretty stupid strategy that makes everything worse for you and everyone else"
X Link 2025-12-01T17:21Z 431.1K followers, 28.9K engagements
"My health insurance in the US. Family of X. $15000 deductible. $20000 out-of-pocket. You can't call this health insurance. This is just to protect against a catastrophe. We are the dumbest country in the world that can't figure out how to fix this broken system"
X Link 2025-12-01T22:02Z 431K followers, 661.9K engagements
"Prompt engineering is dead. Context engineering is the new king. I've talked to many people building agents and most are duct-taped together using a bunch of services: One to store messages One to manage memory Artifacts in S3 A database acting as a journal This is a complex mess. Agents don't learn from their experience. They can't observe their own behavior and do something with it because it's all scattered everywhere. Fortunately we are trying to solve this: Check out Acontext (GitHub Repository link below). This is the first platform I've seen that's centered around context: STORE what"
X Link 2025-12-10T16:02Z 431K followers, 22.4K engagements
"Reviewing AI-generated code is the new bottleneck. Writing code was slow before. Now in a few minutes you can generate virtually unlimited lines of code using AI. But developers are now spending XX% of their time checking that code. That's the new bottleneck. I've been using Cline and they have a feature that actually helps with this: You can click a button and have Cline generate an inline explanation for every single line of AI-generated code. Take a look at the attached video. What it does: X. You can read the code and the explanation together X. You don't lose context while reviewing code"
X Link 2025-12-04T16:01Z 431K followers, 63.4K engagements
"What a stupid timeline we are living in. Beyond parody"
X Link 2025-12-05T19:37Z 431K followers, 47.5K engagements
"You don't need another tutorial. You don't learn football by watching videos or reading a rule book. You learn by running and kicking the ball. Stop bookmarking webpages and downloading books. Writing code and building a project is the only way to learn"
X Link 2025-12-07T13:15Z 431.1K followers, 19.2K engagements
"My mom is obsessed with AI. She has taken her entire photo catalog and generated 1000 variations of all of the family photos using Dreamina + Nano Banana Pro. Probably one of the best Christmas gifts you can give someone is to show them how to do this"
X Link 2025-12-09T20:05Z 431K followers, 12.1K engagements
"If you can't see what an agent does you can't improve it you can't debug it and you can't trust it. It's crazy how many teams are building agents with no way to understand what they're doing. Literally ZERO observability. This is probably one of the first questions I ask every new team I meet: Can you show me the traces of a few executions of your agents Nada. Zero. Nilch. Large language models make bad decisions all the time. Agents fail and you won't realize it until somebody complains. At a minimum every agent you build should produce traces showing the full request flow latency analysis"
X Link 2025-12-10T14:01Z 431K followers, 14.4K engagements
"it sure looks like google is winning"
X Link 2025-12-10T18:07Z 431K followers, 22.1K engagements
"This is insane power for software developers: An agentic principal engineer that you can install as an MCP server. It can index any codebase and surface insights you didn't even know about. This is likely the most useful MCP server you'll start using. It's called @Qodo Aware and it's the first production-ready deep research agent built for navigating and understanding enterprise-scale codebases. Find the GitHub repository in a comment below. The beauty of this is that you don't have to switch tools to take advantage of it: you can keep using Cursor Visual Studio Code Claude Code or whatever"
X Link 2025-09-10T19:10Z 431K followers, 78.6K engagements
"I paid $150/mo for a mobile developer. That was around X years ago. Today you can describe the app you want to create in plain English and have AI build it for you. Check out the video I recorded using OnSpace AI. It took me XX minutes to build a mobile application that scans receipts and uses OCR to read all of the information. A few of the highlights: You can create what you want by just describing it. You can plug in cutting-edge AI models instantly. No API keys no setup. You can stitch together reasoning workflows data and UI without writing a single line of code. Whatever you build runs"
X Link 2025-11-27T16:02Z 431K followers, 45.1K engagements
"There is a massive opportunity right now for anyone who can walk into a company and automate the grunt work they do every day. It doesn't matter what programming language you use or libraries you like or whether you know what micro-services do or if you like Claude Code or Cursor. No one cares about any of that. If you can identify a problem and convince them you can solve it you'll make a killing"
X Link 2025-11-28T13:45Z 431K followers, 49.8K engagements
"Running LLMs at scale is just too expensive. This is a common concern for most companies I talk to. We need faster and cheaper models. Here is a FREE session where the Nebius team will break down LLM distillation and give you an end-to-end practical guide on how to use it. You'll learn: How to train smaller models that stay fast and accurate How to cut inference costs without painful trade-offs How to deploy distilled models in production with SLAs and zero-retention guarantees For those of you who are hands-on you'll get a full demo that will walk you through the entire workflow. If you're"
X Link 2025-12-01T20:05Z 431K followers, 15.6K engagements
"I used to charge $20000 per week to come and help your team fix your software architecture and revive your failed projects. A big part of that was finding and fixing bugs. It's mind-blowing to realize you can now do a lot of the same work for $XXX. Here is a tip: Check out Max at Max is an "AI software engineer" that does something I haven't seen done yet: X. It helps you build your application X. The agent then loads and uses the app autonomously to find and fix bugs 👀 I think this is new and it's got huge potential Max loads your app clicks around types and uses it to test it. This is"
X Link 2025-12-03T20:15Z 431.1K followers, 31.8K engagements
"To me open source has always been about "opening the source code." That's literally been my understanding since forever. Well it turns out that somebody decided that "open source" doesn't mean that and instead they have a purity test to define whether you can call something "open" or not. If you subscribe to this redefinition of the term I'm happy for you. But you are absolutely nobody to tell me what open source is and how and when I can use the term"
X Link 2025-12-05T17:17Z 431K followers, 13.5K engagements
"Here is what I like about Google ADK to build code-first agents: X. Open-source X. Backed by Google X. Model-agnostic - It works with Gemini OpenAI models Claude Llama or whatever other model you want to use. X. Multi-agent by design - You can start with a simple agent and evolve your system to as many specialized agents as you need. X. Code-first - This is all Python and you have complete control over everything (instead of depending on a tool that hides the code behind a drag & drop interface.) X. You get a CLI and a visual web UI for debugging and evaluation. X. Enterprise-ready from day 1"
X Link 2025-12-05T13:25Z 431K followers, 16.9K engagements
"We have X new benchmark datasets to evaluate LLMs' structured-output capabilities. (Because existing benchmarks are full or errors) X. Data Table Analysis - Evaluates how well an LLM extracts key metadata about each table. X. Insurance Claims Extraction - Evaluates how well an LLM extracts information about each claim. X. Financial Entities Extraction - Evaluates how well an LLM extracts structured financial and contextual entities mentioned in a document. X. PII Extraction - Evaluates how well an LLM extracts different categories of PII from the document. This is awesome"
X Link 2025-12-05T17:58Z 431K followers, 11.3K engagements
"This is still the way I recommend most people start with machine learning: X. Start with Python X. Learn to use Google Colab X. Take a Pandas tutorial X. Then a Seaborn tutorial X. Learn how to use Decision Trees X. Finish Kaggle's "Intro to Machine Learning" X. Solve the Titanic challenge"
X Link 2025-10-19T15:15Z 431K followers, 461.6K engagements
"Best GPU prices you'll find out there: H100 GPU for $2.00/hr H200 GPU for $2.30/hr Even better: You can literally move all of your data to Nebius for free including a free PoC with support and get up to X months for free Nebius is collaborating with me on this post and their deal is just amazing This is insanely cheap and you should check it out if you are looking to fine-tune models run experiments or do some production testing. Check this out:"
X Link 2025-09-22T15:10Z 431K followers, 59.2K engagements
"Web scraping is a critical skill for software developers At a time when every large company is training large language models learning how to scrape content from the web is a top skill to learn. Web scraping Collecting data at scale and storing it in a structured format like CSV or JSON. Here is how the process works at a high level: You send a request to the URL you want to scrape The server responds with the HTML of the page Your code parses the HTML and collects the data Rinse and repeat for every URL from which you want to scrape data. Here is a free and open-source tool you can use:"
X Link 2025-09-30T15:15Z 431.1K followers, 154.1K engagements
"Ive been building software for XX years more than XX professionally. I understand architecture principles and best practices. No you cant vibe-code Salesforce SAP or any other complex enterprise app"
X Link 2025-11-26T19:22Z 431K followers, 372.8K engagements
"I don't believe you can vibe-code complex applications. Put another way your ability to vibe-code software is inversely proportional to its complexity. Vibe-coding is like going into the "Minotaur Labyrinth" without Theseus' thread: you'll invariably get lost. More accurately you'll be stuck in an endless loop going in circles with the model: Things that used to work stop working Your code starts bloating Solutions become over-engineered And all of this will happen while working with popular tech where models have plenty of training. Try to step outside the beaten path (use anything a bit"
X Link 2025-11-27T13:10Z 431K followers, 77.7K engagements
"Imagine getting mad because @37signals opened the source code of their newest product but kept the right to monetize it. I will never understand people like this. If they kept the source code closed (like literally every single saas builder does) nobody would have complained. But they dared to open-source that code so that's a problem somehow @37signals is a gift to our industry. We should have more companies doing as they do not fewer"
X Link 2025-12-04T17:34Z 431.1K followers, 65.5K engagements
"Kiro keeps getting better For those of you who don't know Kiro is the AI-powered coding assistant from AWS. They just released "Kiro Powers" and here is how it works: Instead of dumping every MCP tool into context and hoping your agent figures it out this feature helps Kiro load exactly what you need exactly when you need it. The whole idea is to prevent context rot and prevent having to create rules and custom instructions by hand. You now have a POWER .md file that acts as the onboarding manual the agent reads. It tells the agent what tools to use when to use them and how to use them"
X Link 2025-12-05T16:10Z 431K followers, 12.4K engagements
"I think it's time to start talking about AgentOps. DevOps MLOps AgentOps If you want autonomous agents that work and scale we need to start formalizing the discipline that supports them. Some of the things everyone has to worry about: Agent evaluations (using LLM-as-a-judge or any other technique) Metrics-driven releases (ship new versions when they beat the threshold) A/B testing agents canary releases shadow deployments Enabling observability and tracing Implementing and integrating human feedback loops You can't build a reliable agent without worrying about everything above"
X Link 2025-12-06T14:01Z 431K followers, 27.7K engagements
"Stripe wanted to buy this company but the founder declined their $XXX billion offer. Today the global financial system is about to change because of what Airwallex has been doing. Now they are a $1B ARR global fintech and they're building what could be the first execution layer for autonomous finance. These AI agents can now do what used to take hours: Pull receipts from emails Slack WhatsApp Match them to transactions Categorize automatically Catch policy exceptions before they slip through The underlying infrastructure they are using is pretty impressive: XX% of payments run on local rails"
X Link 2025-12-08T14:56Z 431K followers, 40.6K engagements
"This story is incredible: Back in 2005 a guy switched to Computer Science because he wanted to work at Google. After a ton of work he finally got an internship at the company. When he left he founded his first company. Google backed it. He then built something far more ambitious: Emergent an end-to-end agentic vibe-coding platform. And he just closed funding from Google's AI Futures Fund As of today Emergent has: 2.5M+ people using it $25M ARR (they got there in X months) 3M+ apps shipped You can use Emergent to create fully functioning products with a prompt: X. The front-end application 2."
X Link 2025-12-10T20:10Z 431K followers, 137.4K engagements
/creator/twitter::svpino