kibh

Developer. Freelance
Building products for clients.

[kibh @tprototypex]

wrote a blog about "How AI generates text?"
https://t.co/lC93ZiAzaT

[kibh @tprototypex]

your gut feeling is not random

its your taste, experience and choices which build you up from past

[kibh @tprototypex]

have something as your thing
like “i don’t do this particular thing”

as “your rule”

ts helps u in long run, a self control

[kibh @tprototypex]

@rosmine i thought if we could train on our own writing pieces for making it sound more written by oneself

but yeah the training data, will be too less to teach llm

[Kishmattt @spdiLf]

Jay Nepal

[kibh @tprototypex]

@MitcheIl agent

[kibh @tprototypex]

fuck

bad day to end 25

[kibh @tprototypex]

karpathy 🙏

[Ahmad @TheAhmadOsman]

great advice to become an expert at anything from Andrej Karpathy

[kibh @tprototypex]

with every achievement comes new responsibility

[kibh @tprototypex]

me trying to use SLMs without fine tuning.

[Santiago @svpino]

Massive release here!

First, MCP. Then, A2A. Now, we have a new AI protocol.

AG-UI is the Agent-User Interaction Protocol. This is a protocol for building user-facing AI agents. It's a bridge between a backend AI agent and a full-stack application.

Up to this point, most agents are backend automators: form-fillers, summarizers, and schedulers. They are useful as backend tools.

But, interactive agents like Cursor can bring agents to a whole new set of domains, and have been extremely hard to build.

But not anymore!

If you want to build an agent that co-works with users, you need:

• Real-time updates
• Tool orchestration
• Shared mutable state
• Security boundaries
• UI synchronization

AG-UI gives you all of this.

It’s a lightweight, event-streaming protocol (over HTTP/SSE/webhooks) that creates a unified pipe between your agent backend (OpenAI, Ollama, LangGraph, custom code) and your frontend.

Here is how it works:

• Client sends a POST request to the agent endpoint
• Then listens to a unified event stream over HTTP
• Each event includes a type and a minimal payload
• Agents emit events in real-time
• The frontend can react immediately to these events
• The frontend emits events and context back to the agent

Check the link to the protocol in the next post:

[kibh @tprototypex]

System prompt you write helps to understand your personality

[maharshi @maharshii]

if you are an engineer, drop everything you are doing right now and watch Pantheon!

[Kishmattt @spdiLf]

Hello everyone 👋
We just launched this app on the Play Store: https://t.co/ONVAUXBWIP

Would love to know what you think.
If you’re on Apple(coming soon), no worries—feel free to share it or check out our site! (https://t.co/UzX9qSoEsl)

[kibh @tprototypex]

@ionleu https://t.co/wY7R3BZnlO

[kibh @tprototypex]

need to do fine tuning --

one option is rag but sometime it does not understand the overall documents

The other is fine tuning the pretraged weights but this needs a high amount of training data to make it more inclined to "new dataset"

Don't know what's the right middle.

[kibh @tprototypex]

@adxtya_jha i can write some code

[kibh @tprototypex]

@harshbhatt7585 @cneuralnetwork same

[kibh @tprototypex]

tests are running so here i am

[kibh @tprototypex]

@Slothenater Netflix example is removing decision making cause it makes people difficult which is very subtle to decide.

But your another example of ballot bin is adding a decision making...

this two example contradicts

can you please explain this part ??

[kibh @tprototypex]

added a chat feature in Nepshorts.

Now you can query on today's news and get your response. No need to scroll the feed.
#llm #rag #embedding