What is today’s beat?

RELEASES
🧨 Huggingface 2025 roadmap
🧨 App to find out if your image is AI or NOT

NEWSROOM
🗞️ Devin 1.2 Update: AI Engineer Enhances Coding with Smarter Reasoning
🗞️ChatGPT's Head of Product to Testify in US Government's Case Against Google
🗞️ Neura Robotics Secures €120M in Series B for AI-Powered Exoskeletons

BUILDER BYTES
⭐️ Building Knowledge Graph Agents with LlamaIndex Workflows
⭐️ Apache Hudi - Big Data Processing & Storage
⭐️ Facebook research released code for Coconut - Large Language Model Reasoning in Latent Space

COMMUNITY
🤩 A few tools for developers and product managers
🤩 A free Huggingface course on AI agents (sign up link included)

FOLLOW US ON SOCIAL FOR MORE

Your FREE newsletter
 share or subscribe 
to show support

🧨 HUGGING FACE 🧨

MULTI-BACKEND SUPPORT FOR TEXT GENERATION INFERENCE (TGI)

Text Generation Inference (TGI) by Hugging Face now supports multiple backends, enhancing the deployment and performance of large-language models (LLMs) by offering flexibility for users to choose the most optimal inference engine based on their hardware and model requirements. Here is a sneak peak at their 2025 roadmap.

• NVIDIA TensorRT-LLM backend
  tensorRt is Nvidia GPU acceleration!

• Llama.cpp backend
  C++ > python

• vLLM backend
  open source, helps you use what models you want.

• AWS Neuron backend
  runs AI on AWS.

• Google TPU backend
  not GPU (nvidia) but a custom Google ASICs for speed.

If rust makes you hot, you’ll be glad to know they are using it.

Read the full article here

🗞️ NEWSROOM 🗞️ 

What’s hot in tech right now?

Devin 1.2 Update: AI Engineer Enhances Coding with Smarter Reasoning
Devin 1.2 update boosts AI-assisted coding with advanced context reasoning, smarter integrations, and voice commands, streamlining development workflows for more efficient programming.

ChatGPT's Head of Product to Testify in US Government's Case Against Google
ChatGPT's head of product is set to testify in the US government's antitrust case against Google, shedding light on competition in AI-driven search and the evolving role of generative models.

Neura Robotics Secures €120M in Series B for AI-Powered Exoskeletons
Neura Robotics raises €120M to advance its AI-powered exoskeletons, aimed at enhancing mobility and rehabilitation, pushing forward human-robot collaboration in the medical and industrial sectors.

⭐️ TOOL OF THE WEEK ⭐️

AI or NOT AI app

This tool is created by Tecent. Tencent runs China's biggest messaging app, WeChat. The company this year launched an AI chatbot based on its Hunyuan model. It’s kind of like Chinese Facebook + Whatsapp. Don’t worry, they aren’t going to steal your information, and you won’t end up in a bathtub without your liver begging for help. It just analyzes the image and says “AI” or “NOT AI.” If you’ve watched the show Silicon Valley, it’s kind of like the “HOT DOG” or “NOT HOT DOG” app release - check this video out for a funny reference. So, next time you find some garbage on the internet, throw it at this app and see what happens.

TRY IT HERE

⭐️ TUTORIAL ⭐️

What will we learn today?

LangGraph Cloud introduces functionality for scheduling automated tasks, referred to as "Cron Jobs," allowing users to run graphs based on a defined schedule, such as weekly email reminders, without needing to write custom scripts.

Key Takeaways

1. Quick and Easy: Yeah, it’s stupid easy and it will make you look super smart. Just a few lines of code and voila - you’re an AI practitioner!

2. Run on a thread: This example shows how to set up a cron job to run on a thread.

from langgraph_sdk import get_client

client = get_client(url=<DEPLOYMENT_URL>)
# Using the graph deployed with the name "agent"
assistant_id = "agent"
# create thread
thread = await client.threads.create()
print(thread)

# This schedules a job to run at 15:27 (3:27PM) every day
cron_job = await client.crons.create_for_thread(
    thread["thread_id"],
    assistant_id,
    schedule="27 15 * * *",
    input={"messages": [{"role": "user", "content": "What time is it?"}]},
)

await client.crons.delete(cron_job["cron_id"])

Have a product and want to contribute?
Or, want to learn a topic?
👉️ tell us

⭐️ BUILDER BYTES ⭐️

What’s hot for builders right now?

BLOG | Building Knowledge Graph Agents with LlamaIndex Workflows

Explores how LlamaIndex Workflows can improve text2cypher tasks by integrating structured data retrieval from graph databases with multi-step agentic strategies, boosting accuracy and reliability.

PROJECT | Apache Hudi - Big Data Processing & Storage

An open-source APACHE data lake development finally released 1.0! It helps with efficient management of data lakes with features like versioning, indexing, and schema evolution.

CODE | Facebook research released code for Coconut - Large Language Model Reasoning in Latent Space

Coconut - large language reasoning in latent space - was gained quick popularity in the community. They have now released its code - gives models ability to reason within a continuous latent space focusing on tasks that require multi-step reasoning and structured thought processes.

🤩 COMMUNITY 🤩

What’s the latest beat?

THANK YOU

Looking to promote your company, product, service, or event?