Harshi

Machine Learning Guy

[Harshi @Harshit9105]

If you don't know me yet, I create content around:

🐍 Python
🤖 Machine Learning
📚 NLP → Mostly LLMs
🛠 MLOps

✔️ Follow me @Harshit9105
to stay on top of the above topics.

[Akshay 🚀 @akshay_pachaar]

You can now deploy any ML model, RAG, or Agent as an MCP server.

And it takes just 10 lines of code.

Here's a step-by-step breakdown (100% private):

[Harshi @Harshit9105]

I’m using @grok and its pretty amazing. i hope @grok will see this post and if they comment I will take the premium of x

[Avi Chawla @_avichawla]

Fine-tune 100+ LLMs directly from a UI!

LLaMA-Factory lets you train and fine-tune open-source LLMs and VLMs without writing any code.

Supports 100+ models, multimodal fine-tuning, PPO, DPO, experiment tracking, and much more!

100% open-source with 50k stars!

[Avi Chawla @_avichawla]

How LLMs train LLMs, clearly explained (with visuals):

[bhatnaturally  🇮🇳 @bhatnaturally]

Sorry basic question: what do we gain by getting PoK (if at all)

[Vikrant @Vikspeaks1]

Video from Sirsa Haryana. Something big intercepted over there. Mind the sound.

[Mr Sinha @Mrsinha]

Big news : Pakistan had attacked Kutch (Gujarat) using three drones..

The Indian defence system has shot them down successfully...

[Harshi @Harshit9105]

Today will be Pakistan's last night

Jai Hind🇮🇳

[Rajnath Singh @rajnathsingh]

भारत माता की जय!

[M. Faraz (CityNews021) @mfaraz021]

Pakistan has converted "Operation Sindoor" into "Operation Widwaah"
🤣🤣🤣

#PakistanZindabad
#PakistanArmy
#LOC
#LineOfControl
#IndiaPakistanTensions
#IndianFalseFlagExposed
#IndiaPakistan
#IndiaPakWar

@PMOIndia
@majorgauravarya
@ArnabGoswamiRTv

[Gavin Baker @GavinSBaker]

Grok-3 is the first model *ever* to score over 1400 on Chatbot Arena and outperforms the best publicly available reasoning models from OpenAI and Google.

xAI was founded 13 years after Deepmind and 8 years after OpenAI and is now ahead of both. The “SR-71 Blackbird” of AI labs.

[Arena.ai @arena]

BREAKING: @xAI early version of Grok-3 (codename "chocolate") is now #1 in Arena! 🏆

Grok-3 is:
- First-ever model to break 1400 score!
- #1 across all categories, a milestone that keeps getting harder to achieve

Huge congratulations to @xAI on this milestone! View thread 🧵 for more insights into Grok-3's performance after ~8K votes in the Arena.

[Poonam Soni @CodeByPoonam]

DeepSeek R1 just broke the Internet, and people are going crazy over it.

SPOILER: ChatGPT is now falling behind.

13 WILD examples so far (Don't miss the 5th one):

[Poonam Soni @CodeByPoonam]

Goodbye ChatGPT

It’s only been 5 days since Deepseek R1 dropped, and the World is already blown away by its potential.

13 examples that will blow your mind (Don't miss the 5th one):

[Bindu Reddy @bindureddy]

2025 Will Be The Year of AI Agents

Organizations will have 50 - 500 agents that automate various tasks.

These AI agents will

- Talk and perform actions on Enterprise systems
- Automate workflows
- Have an understanding of Enterprise data
- Autonomously perform tasks on schedule or triggers
- Create documents, respond to emails, and chat with you on Slack

A lot of knowledge work will be automated!

[Alexandr Wang @alexandr_wang]

Since ChatGPT dropped in 2022, AI progress has been dramatic.

But it's also been predictable—new models, bigger chip clusters, more chatbots.

Not in 2025.

Here are the three big changes to watch for over the next 12 months 🧵

1/8

[edwin @edwinarbus]

how it started how it's going 🍓

[Quinton Newman @QFromDaaS]

WE DID IT. WE FINALLY DID IT 😭😭😭😭

@sama @OpenAI We're so proud 🥹 #AGI acheived??

#GPT5 #o1preview #gpto1

[Rohan Paul @rohanpaul_ai]

"LLMs Will Always Hallucinate, and We Need to Live With This" - 🤔🤔

Key points from the paper. 👇

🧠 Hallucinations in LLMs not just mistakes, but inherent property. Arise from undecidable problems in training and usage process. Can't be fully eliminated through architectural improvements or data cleaning.

🔬 They use computational theory and Gödel's incompleteness theorems to explain hallucinations. Argue that LLM structure inherently leads to some inputs causing model to generate false or nonsensical information.

🚫 Complete elimination of hallucinations impossible due to undecidable problems in LLM foundations. No amount of tweaks or fact-checking can fully solve this issue. Fundamental limitation of current LLM approach.

------

🧮 Gödel's incompleteness theorems:

👉 First theorem: Any consistent formal system powerful enough to encode arithmetic contains statements that are true but unprovable within the system.

👉 Second theorem: Such a system cannot prove its own consistency.