Rutam • Gen AI Pro • Freelancer
@rutamstwt.bsky.social
Generative AI Professional & freelance AI expert, I help turn your ideas into real-world AI products and solutions, from agent, chatbots to automations
Slow pip installs mess with CI/CD, but UV makes that process much quicker.
That's a win for devops who are constantly optimizing for speed and resource usage, and that has a big impact on the overall business.
That's a win for devops who are constantly optimizing for speed and resource usage, and that has a big impact on the overall business.
December 14, 2024 at 5:05 PM
Slow pip installs mess with CI/CD, but UV makes that process much quicker.
That's a win for devops who are constantly optimizing for speed and resource usage, and that has a big impact on the overall business.
That's a win for devops who are constantly optimizing for speed and resource usage, and that has a big impact on the overall business.
Built in Rust, UV delivers the kind of performance and reliability we all want, an efficient tool made with performance in mind.
It's not just some script, it's a serious approach to speed; that blazingly fast reputation is well deserved.
It's not just some script, it's a serious approach to speed; that blazingly fast reputation is well deserved.
December 14, 2024 at 5:05 PM
Built in Rust, UV delivers the kind of performance and reliability we all want, an efficient tool made with performance in mind.
It's not just some script, it's a serious approach to speed; that blazingly fast reputation is well deserved.
It's not just some script, it's a serious approach to speed; that blazingly fast reputation is well deserved.
Tired of the nightmare of Python global libraries? UV is addressing that head-on, which is pretty cool.
It's about cutting through all that chaos that we all know too well. A unified tool can only make things much better.
It's about cutting through all that chaos that we all know too well. A unified tool can only make things much better.
December 14, 2024 at 5:05 PM
Tired of the nightmare of Python global libraries? UV is addressing that head-on, which is pretty cool.
It's about cutting through all that chaos that we all know too well. A unified tool can only make things much better.
It's about cutting through all that chaos that we all know too well. A unified tool can only make things much better.
It aims to be the only tool you need to manage Python projects by bundling pip, venv, pipx, and even ruff.
That's less context switching, which results in a much smoother development experience, it's about working smarter, not harder.
That's less context switching, which results in a much smoother development experience, it's about working smarter, not harder.
December 14, 2024 at 5:05 PM
It aims to be the only tool you need to manage Python projects by bundling pip, venv, pipx, and even ruff.
That's less context switching, which results in a much smoother development experience, it's about working smarter, not harder.
That's less context switching, which results in a much smoother development experience, it's about working smarter, not harder.
UV is built as a seamless swap for pip, pip-tools, and virtualenv, it’s a minimal disruption upgrade for your setup, which is always a plus.
This drop-in replacement does not force you to change everything, it's just a lot smoother, you know?
This drop-in replacement does not force you to change everything, it's just a lot smoother, you know?
December 14, 2024 at 5:05 PM
UV is built as a seamless swap for pip, pip-tools, and virtualenv, it’s a minimal disruption upgrade for your setup, which is always a plus.
This drop-in replacement does not force you to change everything, it's just a lot smoother, you know?
This drop-in replacement does not force you to change everything, it's just a lot smoother, you know?
It's about getting back precious time and cutting through the frustration that slows down your workflow.
It's the little things that make a big difference to your productivity.
We all know how long it takes to get started with some projects, right?
It's the little things that make a big difference to your productivity.
We all know how long it takes to get started with some projects, right?
December 14, 2024 at 5:05 PM
It's about getting back precious time and cutting through the frustration that slows down your workflow.
It's the little things that make a big difference to your productivity.
We all know how long it takes to get started with some projects, right?
It's the little things that make a big difference to your productivity.
We all know how long it takes to get started with some projects, right?
Setting up Python projects can often feel like a waiting game, not a coding session, don't you think?
The speed of UV is so fast it might make you wonder if the install actually completed, it really is that quick.
The speed of UV is so fast it might make you wonder if the install actually completed, it really is that quick.
December 14, 2024 at 5:05 PM
Setting up Python projects can often feel like a waiting game, not a coding session, don't you think?
The speed of UV is so fast it might make you wonder if the install actually completed, it really is that quick.
The speed of UV is so fast it might make you wonder if the install actually completed, it really is that quick.
Every minute spent waiting is a minute not creating.
But what if there was a way to instantly speed up your workflow?
But what if there was a way to instantly speed up your workflow?
December 14, 2024 at 5:05 PM
Every minute spent waiting is a minute not creating.
But what if there was a way to instantly speed up your workflow?
But what if there was a way to instantly speed up your workflow?
You can test it for free on Google AI Studio and via the Gemini API. They're giving Free 10 RPM 1500 req/day. I’m thinking about what new projects I could build. What about you?
December 12, 2024 at 12:02 PM
You can test it for free on Google AI Studio and via the Gemini API. They're giving Free 10 RPM 1500 req/day. I’m thinking about what new projects I could build. What about you?
And the best part? Native tool use. Gemini 2.0 Flash can now call tools like Google Search and execute code, along with custom third-party functions, directly. Imagine what we can do! cutting research times down considerably
December 12, 2024 at 12:02 PM
And the best part? Native tool use. Gemini 2.0 Flash can now call tools like Google Search and execute code, along with custom third-party functions, directly. Imagine what we can do! cutting research times down considerably
Let's look at what this means for us developers. Expect better performance across the board: text, code, video, spatial understanding; plus, it's multimodal now, meaning it can output text, audio, and images all from a single API call.
December 12, 2024 at 12:02 PM
Let's look at what this means for us developers. Expect better performance across the board: text, code, video, spatial understanding; plus, it's multimodal now, meaning it can output text, audio, and images all from a single API call.
Gemini 2.0 Flash is here, and it's a significant leap forward.
It like Gemini 1.5 Pro got a huge upgrade , adding a range of new features. The fact that it's twice as fast is very impressive, don't you think?
It like Gemini 1.5 Pro got a huge upgrade , adding a range of new features. The fact that it's twice as fast is very impressive, don't you think?
December 12, 2024 at 12:02 PM
Gemini 2.0 Flash is here, and it's a significant leap forward.
It like Gemini 1.5 Pro got a huge upgrade , adding a range of new features. The fact that it's twice as fast is very impressive, don't you think?
It like Gemini 1.5 Pro got a huge upgrade , adding a range of new features. The fact that it's twice as fast is very impressive, don't you think?
Oh, and there's a visual IDE too! It's super handy for debugging and visualizing your agents. Plus, LangGraph plays nicely with LangChain. It's got integrations with tons of different LLMs, vector stores, and other tools.
December 11, 2024 at 5:05 PM
Oh, and there's a visual IDE too! It's super handy for debugging and visualizing your agents. Plus, LangGraph plays nicely with LangChain. It's got integrations with tons of different LLMs, vector stores, and other tools.
LangGraph has features that make all this possible: persistence, streaming, human-in-the-loop capabilities, and advanced controllability. Persistence helps with memory, streaming keeps things flowing, human-in-the-loop lets you, well, bring humans into the process.
December 11, 2024 at 5:05 PM
LangGraph has features that make all this possible: persistence, streaming, human-in-the-loop capabilities, and advanced controllability. Persistence helps with memory, streaming keeps things flowing, human-in-the-loop lets you, well, bring humans into the process.
LangGraph represents these workflows as graphs. In these graphs, nodes are the steps in your application – a tool call, a retrieval step, anything like that. Edges are simply the connections between these nodes. You've got flexibility in how you set up these nodes and edges.
December 11, 2024 at 5:05 PM
LangGraph represents these workflows as graphs. In these graphs, nodes are the steps in your application – a tool call, a retrieval step, anything like that. Edges are simply the connections between these nodes. You've got flexibility in how you set up these nodes and edges.
This is where LangGraph steps in. The big idea behind LangGraph is to let you build agents that are both flexible and reliable. It's about getting the best of both worlds. How? By letting you mix and match developer control with LLM control.
December 11, 2024 at 5:05 PM
This is where LangGraph steps in. The big idea behind LangGraph is to let you build agents that are both flexible and reliable. It's about getting the best of both worlds. How? By letting you mix and match developer control with LLM control.
But, you know what? There's a catch. The more control you give an LLM, the less reliable it tends to be. It's like the difference between a well-tested function and a brand-new, experimental one. You're trading reliability for flexibility.
December 11, 2024 at 5:05 PM
But, you know what? There's a catch. The more control you give an LLM, the less reliable it tends to be. It's like the difference between a well-tested function and a brand-new, experimental one. You're trading reliability for flexibility.
Now, here's where it gets interesting. There are different types of agents, kind of like a spectrum of how much control you give the LLM. On one end, you have "routers." These are simple: the LLM chooses between a few options at a single step.
December 11, 2024 at 5:05 PM
Now, here's where it gets interesting. There are different types of agents, kind of like a spectrum of how much control you give the LLM. On one end, you have "routers." These are simple: the LLM chooses between a few options at a single step.
But what if you want your LLM to be a bit more... adaptable? That's where agents come in. An agent is an LLM that defines its own workflow. So, chains are fixed, developer-defined workflows, while agents are LLM-defined workflows.
December 11, 2024 at 5:05 PM
But what if you want your LLM to be a bit more... adaptable? That's where agents come in. An agent is an LLM that defines its own workflow. So, chains are fixed, developer-defined workflows, while agents are LLM-defined workflows.
So, how do we make LLMs more useful? You've probably heard the term "chain" thrown around. A chain is a set of steps that happen before and after an LLM does its thing. Chains are great because they're reliable. You set them up, and they do the same thing every time.
December 11, 2024 at 5:05 PM
So, how do we make LLMs more useful? You've probably heard the term "chain" thrown around. A chain is a set of steps that happen before and after an LLM does its thing. Chains are great because they're reliable. You set them up, and they do the same thing every time.
First off, what's the deal with language models on their own? Well, they're a bit limited, aren't they? Think about it: a solo language model, or LLM, is a brain without hands. They can't use tools, access external data like docs, or even manage multi-step workflows on their own.
December 11, 2024 at 5:05 PM
First off, what's the deal with language models on their own? Well, they're a bit limited, aren't they? Think about it: a solo language model, or LLM, is a brain without hands. They can't use tools, access external data like docs, or even manage multi-step workflows on their own.
Let's talk about LangGraph. @langchain.bsky.social This is going to be a multi-part series, a thread 🧵, breaking down what LangGraph is all about and why you, as a developer, should be excited about it.
December 11, 2024 at 5:05 PM
Let's talk about LangGraph. @langchain.bsky.social This is going to be a multi-part series, a thread 🧵, breaking down what LangGraph is all about and why you, as a developer, should be excited about it.
I used to name my git commits something like “minor changes” or even stupid i used to use the same commit message as my last commit and thought that i might rebase later but never did!
December 10, 2024 at 11:50 AM
I used to name my git commits something like “minor changes” or even stupid i used to use the same commit message as my last commit and thought that i might rebase later but never did!
The AI/LLM agents stack represents a major shift from the traditional LLM stack.
The primary distinction lies in how state is managed: while LLM serving platforms are typically stateless, agent serving platforms must be stateful, maintaining the agent's state on the server side.
The primary distinction lies in how state is managed: while LLM serving platforms are typically stateless, agent serving platforms must be stateful, maintaining the agent's state on the server side.
December 10, 2024 at 11:39 AM
The AI/LLM agents stack represents a major shift from the traditional LLM stack.
The primary distinction lies in how state is managed: while LLM serving platforms are typically stateless, agent serving platforms must be stateful, maintaining the agent's state on the server side.
The primary distinction lies in how state is managed: while LLM serving platforms are typically stateless, agent serving platforms must be stateful, maintaining the agent's state on the server side.
INVOICE MANAGEMENT JUST GOT AN UPGRADE 🚀
I used to spend HOURS manually extracting data from invoices. 🤯 Not anymore!
AutoExtract the AI invoice manager, which uses Google Gemini AI to extract information from any format—Excel, PDF, or Images.
Check it out:
#ai #openai #google #langchain
I used to spend HOURS manually extracting data from invoices. 🤯 Not anymore!
AutoExtract the AI invoice manager, which uses Google Gemini AI to extract information from any format—Excel, PDF, or Images.
Check it out:
#ai #openai #google #langchain
November 28, 2024 at 9:19 AM
INVOICE MANAGEMENT JUST GOT AN UPGRADE 🚀
I used to spend HOURS manually extracting data from invoices. 🤯 Not anymore!
AutoExtract the AI invoice manager, which uses Google Gemini AI to extract information from any format—Excel, PDF, or Images.
Check it out:
#ai #openai #google #langchain
I used to spend HOURS manually extracting data from invoices. 🤯 Not anymore!
AutoExtract the AI invoice manager, which uses Google Gemini AI to extract information from any format—Excel, PDF, or Images.
Check it out:
#ai #openai #google #langchain