Taras Trishchuk
@trishchuk.bsky.social
🚀 Meet Tari (XTM), a new layer-1 chain from ex-Monero devs, prioritizing privacy, crypto mining on your Mac/PC.
🔹 Tari Universe app offers a visual
🔹 Growing community with 700k+ airdrop participants.
🔹 Long-term miner incentives with a 1% tail emission after 12 years.
🔹 Tari Universe app offers a visual
🔹 Growing community with 700k+ airdrop participants.
🔹 Long-term miner incentives with a 1% tail emission after 12 years.
May 8, 2025 at 2:41 PM
🚀 Meet Tari (XTM), a new layer-1 chain from ex-Monero devs, prioritizing privacy, crypto mining on your Mac/PC.
🔹 Tari Universe app offers a visual
🔹 Growing community with 700k+ airdrop participants.
🔹 Long-term miner incentives with a 1% tail emission after 12 years.
🔹 Tari Universe app offers a visual
🔹 Growing community with 700k+ airdrop participants.
🔹 Long-term miner incentives with a 1% tail emission after 12 years.
🔧 The MCP server powers tools like Claude Desktop & Cursor by serving GitHub repo data — summaries, file trees, content, metadata, and more. Features include local cloning, caching, autocomplete, and a web UI. #devtools #GitHub #AIdev
www.npmjs.com/package/@tri...
www.npmjs.com/package/@tri...
@trishchuk/github-explorer-mcp
MCP server that provides Github repository info including file content, directory structure and other metadata. Latest version: 0.2.0, last published: 3 days ago. Start using @trishchuk/github-explore...
www.npmjs.com
April 6, 2025 at 4:48 PM
🔧 The MCP server powers tools like Claude Desktop & Cursor by serving GitHub repo data — summaries, file trees, content, metadata, and more. Features include local cloning, caching, autocomplete, and a web UI. #devtools #GitHub #AIdev
www.npmjs.com/package/@tri...
www.npmjs.com/package/@tri...
Monte Carlo Tree Search (MCTS), a method initially developed for game AI, is now used to enhance mathematical reasoning in AI models.
January 10, 2025 at 5:08 PM
Monte Carlo Tree Search (MCTS), a method initially developed for game AI, is now used to enhance mathematical reasoning in AI models.
CAG (Context-Augmented Generation) offers an alternative to the traditional RAG approach. Instead of spending time searching for documents in real time, all necessary information is preloaded into the model's context.
January 10, 2025 at 1:31 PM
CAG (Context-Augmented Generation) offers an alternative to the traditional RAG approach. Instead of spending time searching for documents in real time, all necessary information is preloaded into the model's context.
A minimalistic AI-powered search engine that helps you find information on the internet. Powered by Vercel AI SDK! Search with models like Grok 2.0
github.com/zaidmukaddam...
github.com/zaidmukaddam...
GitHub - zaidmukaddam/miniperplx: A minimalistic AI-powered search engine that helps you find information on the internet. Powered by Vercel AI SDK! Search with models like Grok 2.0.
A minimalistic AI-powered search engine that helps you find information on the internet. Powered by Vercel AI SDK! Search with models like Grok 2.0. - zaidmukaddam/miniperplx
github.com
January 5, 2025 at 8:10 PM
A minimalistic AI-powered search engine that helps you find information on the internet. Powered by Vercel AI SDK! Search with models like Grok 2.0
github.com/zaidmukaddam...
github.com/zaidmukaddam...
Oldies but goldies: B Poljak, Some methods of speeding up the convergence of iteration methods, 1964. Introduces momentum methods to speed up gradient descent schemes.
distill.pub/2017/momentum/
distill.pub/2017/momentum/
Why Momentum Really Works
We often think of optimization with momentum as a ball rolling down a hill. This isn't wrong, but there is much more to the story.
distill.pub
January 5, 2025 at 7:23 PM
Oldies but goldies: B Poljak, Some methods of speeding up the convergence of iteration methods, 1964. Introduces momentum methods to speed up gradient descent schemes.
distill.pub/2017/momentum/
distill.pub/2017/momentum/
Cool collection of OpenAI o1 related papers, blogs, and projects, focussing on reasoning techniques.
📌 Curated papers, blogs, talks, and Twitter discussions about OpenAI's o1 and LLM reasoning
📌 Tracks frontier developments in LLM reasoning capabilities and techniques
bit.ly/4fDRzo0
📌 Curated papers, blogs, talks, and Twitter discussions about OpenAI's o1 and LLM reasoning
📌 Tracks frontier developments in LLM reasoning capabilities and techniques
bit.ly/4fDRzo0
GitHub - hijkzzz/Awesome-LLM-Strawberry: A collection of LLM papers, blogs, and projects, with a focus on OpenAI o1 🍓 and reasoning techniques.
A collection of LLM papers, blogs, and projects, with a focus on OpenAI o1 🍓 and reasoning techniques. - hijkzzz/Awesome-LLM-Strawberry
bit.ly
December 31, 2024 at 12:44 PM
Cool collection of OpenAI o1 related papers, blogs, and projects, focussing on reasoning techniques.
📌 Curated papers, blogs, talks, and Twitter discussions about OpenAI's o1 and LLM reasoning
📌 Tracks frontier developments in LLM reasoning capabilities and techniques
bit.ly/4fDRzo0
📌 Curated papers, blogs, talks, and Twitter discussions about OpenAI's o1 and LLM reasoning
📌 Tracks frontier developments in LLM reasoning capabilities and techniques
bit.ly/4fDRzo0
Transform any document into LLM-ready format for building AI Agents and RAG apps.
With just three lines of Python Code.
100% Opensource.
github.com/DS4SD/docling
With just three lines of Python Code.
100% Opensource.
github.com/DS4SD/docling
GitHub - DS4SD/docling: Get your documents ready for gen AI
Get your documents ready for gen AI. Contribute to DS4SD/docling development by creating an account on GitHub.
github.com
December 22, 2024 at 3:52 PM
Transform any document into LLM-ready format for building AI Agents and RAG apps.
With just three lines of Python Code.
100% Opensource.
github.com/DS4SD/docling
With just three lines of Python Code.
100% Opensource.
github.com/DS4SD/docling
Sparrow is an open-source solution for extracting and processing data from documents and images using ML, handling forms, invoices, receipts, and other structured data with modules for OCR, ML models, and data labeling
github.com/katanaml/spa...
github.com/katanaml/spa...
December 12, 2024 at 7:53 PM
Sparrow is an open-source solution for extracting and processing data from documents and images using ML, handling forms, invoices, receipts, and other structured data with modules for OCR, ML models, and data labeling
github.com/katanaml/spa...
github.com/katanaml/spa...
Quanta has written popularly about this result and quantum error correction in general
www.quantamagazine.org/quantum-comp...
www.quantamagazine.org/quantum-comp...
Quantum Computers Cross Critical Error Threshold | Quanta Magazine
In a first, researchers have shown that adding more “qubits” to a quantum computer can make it more resilient. It’s an essential step on the long road to practical applications.
www.quantamagazine.org
December 10, 2024 at 9:15 PM
Quanta has written popularly about this result and quantum error correction in general
www.quantamagazine.org/quantum-comp...
www.quantamagazine.org/quantum-comp...
A study reveals that zeroing a single “super weight” in an LLM can drastically degrade its performance. The authors propose a method for identifying such weights in one forward pass.
Article: bit.ly/49tUDBr
Code: bit.ly/4g4L8Lr
Article: bit.ly/49tUDBr
Code: bit.ly/4g4L8Lr
The Super Weight in Large Language Models
Recent works have shown a surprising result: a small fraction of Large Language Model (LLM) parameter outliers are disproportionately important to the quality of the model. LLMs contain billions of pa...
arxiv.org
December 9, 2024 at 10:51 PM
A study reveals that zeroing a single “super weight” in an LLM can drastically degrade its performance. The authors propose a method for identifying such weights in one forward pass.
Article: bit.ly/49tUDBr
Code: bit.ly/4g4L8Lr
Article: bit.ly/49tUDBr
Code: bit.ly/4g4L8Lr
Willow has state-of-the-art performance across several metrics, enabling two major achievements.
1. Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
bit.ly/4g4JQAp
1. Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
bit.ly/4g4JQAp
Meet Willow, our state-of-the-art quantum chip
Our new quantum chip demonstrates error correction and performance that paves the way to a useful, large-scale quantum computer.
bit.ly
December 9, 2024 at 10:45 PM
Willow has state-of-the-art performance across several metrics, enabling two major achievements.
1. Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
bit.ly/4g4JQAp
1. Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
bit.ly/4g4JQAp
Ollama now supports structured outputs making it possible to constrain a model’s output to a specific format defined by a JSON schema.
- Parsing data from docs
- Extracting data from images
- Structuring all language model responses
- More reliability and consistency than JSON mode
bit.ly/49ov6JQ
- Parsing data from docs
- Extracting data from images
- Structuring all language model responses
- More reliability and consistency than JSON mode
bit.ly/49ov6JQ
Structured outputs · Ollama Blog
Ollama now supports structured outputs making it possible to constrain a model's output to a specific format defined by a JSON schema. The Ollama Python and JavaScript libraries have been updated to s...
ollama.com
December 6, 2024 at 7:51 PM
Ollama now supports structured outputs making it possible to constrain a model’s output to a specific format defined by a JSON schema.
- Parsing data from docs
- Extracting data from images
- Structuring all language model responses
- More reliability and consistency than JSON mode
bit.ly/49ov6JQ
- Parsing data from docs
- Extracting data from images
- Structuring all language model responses
- More reliability and consistency than JSON mode
bit.ly/49ov6JQ
The universe is quite literally a song
December 6, 2024 at 10:47 AM
The universe is quite literally a song
Reposted by Taras Trishchuk
We're excited to announce our integration with the Model Context Protocol (MCP), a new open protocol introduced by @anthropic.com that "enables seamless integration between LLM applications and external data sources and tools". Check out the full post below: blog.continue.dev/model-contex...
Model Context Protocol x Continue
Last week, Anthropic introduced the Model Context Protocol (MCP), a new open protocol that "enables seamless integration between LLM applications and external data sources and tools". Also last week, ...
blog.continue.dev
December 5, 2024 at 7:54 PM
We're excited to announce our integration with the Model Context Protocol (MCP), a new open protocol introduced by @anthropic.com that "enables seamless integration between LLM applications and external data sources and tools". Check out the full post below: blog.continue.dev/model-contex...
It has come to my attention that many of you are unsure what an AI agent is.
Here’s a simple diagram comparing an AI agent (in its most basic form) to an API call with a tool.
Here’s a simple diagram comparing an AI agent (in its most basic form) to an API call with a tool.
December 5, 2024 at 10:15 PM
It has come to my attention that many of you are unsure what an AI agent is.
Here’s a simple diagram comparing an AI agent (in its most basic form) to an API call with a tool.
Here’s a simple diagram comparing an AI agent (in its most basic form) to an API call with a tool.
December 5, 2024 at 10:05 PM
In the 1990s, the Cypherpunks emerged, advocating for digital privacy through encryption. Among them was Hal Finney, a key figure in Bitcoin’s development. The Cypherpunks created tools like PGP for secure communication and inspired technologies like blockchain and cryptocurrencies.
December 5, 2024 at 6:20 PM
In the 1990s, the Cypherpunks emerged, advocating for digital privacy through encryption. Among them was Hal Finney, a key figure in Bitcoin’s development. The Cypherpunks created tools like PGP for secure communication and inspired technologies like blockchain and cryptocurrencies.
Snowflake's frontier embedding model. Arctic Embed 2.0 adds multilingual support without sacrificing English performance or scalability.
Uncompromising quality for English and non-English retrieval
ollama pull snowflake-arctic-embed2
Uncompromising quality for English and non-English retrieval
ollama pull snowflake-arctic-embed2
December 5, 2024 at 5:46 PM
Snowflake's frontier embedding model. Arctic Embed 2.0 adds multilingual support without sacrificing English performance or scalability.
Uncompromising quality for English and non-English retrieval
ollama pull snowflake-arctic-embed2
Uncompromising quality for English and non-English retrieval
ollama pull snowflake-arctic-embed2
At the end of the Roman Empire, just as at the end of the Russian Empire, inflation spiraled out of control. Enormous government expenditures required endless "printing" of money — until an entire bag of coins couldn't even buy a wheat bag.
November 26, 2024 at 9:19 PM
At the end of the Roman Empire, just as at the end of the Russian Empire, inflation spiraled out of control. Enormous government expenditures required endless "printing" of money — until an entire bag of coins couldn't even buy a wheat bag.
🏄SurfSense
Personal AI Assistant for Internet Surfers and Researchers. Research & Never forget anything you see on the Internet
Fully open source! Check out the code here
github.com/MODSetter/Su...
Personal AI Assistant for Internet Surfers and Researchers. Research & Never forget anything you see on the Internet
Fully open source! Check out the code here
github.com/MODSetter/Su...
November 22, 2024 at 1:40 PM
🏄SurfSense
Personal AI Assistant for Internet Surfers and Researchers. Research & Never forget anything you see on the Internet
Fully open source! Check out the code here
github.com/MODSetter/Su...
Personal AI Assistant for Internet Surfers and Researchers. Research & Never forget anything you see on the Internet
Fully open source! Check out the code here
github.com/MODSetter/Su...
It seems like everyone's number one question with building with AI is "How do you make it not shit?"
The answer is evals. They're like unit tests but for probabilistic systems.
Here's an imaginary API to explain how they work:
The answer is evals. They're like unit tests but for probabilistic systems.
Here's an imaginary API to explain how they work:
November 20, 2024 at 12:29 PM
It seems like everyone's number one question with building with AI is "How do you make it not shit?"
The answer is evals. They're like unit tests but for probabilistic systems.
Here's an imaginary API to explain how they work:
The answer is evals. They're like unit tests but for probabilistic systems.
Here's an imaginary API to explain how they work:
🔗 Agent Protocol: interoperability for language agents
LangGraph is a multi-agent framework. It can communicate with agents defined in other frameworks and languages.
📱 A standard interface for agent communication (Agent Protocol): github.com/langchain-ai...
Blog: blog.langchain.dev/agent-protoc...
LangGraph is a multi-agent framework. It can communicate with agents defined in other frameworks and languages.
📱 A standard interface for agent communication (Agent Protocol): github.com/langchain-ai...
Blog: blog.langchain.dev/agent-protoc...
GitHub - langchain-ai/agent-protocol
Contribute to langchain-ai/agent-protocol development by creating an account on GitHub.
github.com
November 19, 2024 at 8:11 PM
🔗 Agent Protocol: interoperability for language agents
LangGraph is a multi-agent framework. It can communicate with agents defined in other frameworks and languages.
📱 A standard interface for agent communication (Agent Protocol): github.com/langchain-ai...
Blog: blog.langchain.dev/agent-protoc...
LangGraph is a multi-agent framework. It can communicate with agents defined in other frameworks and languages.
📱 A standard interface for agent communication (Agent Protocol): github.com/langchain-ai...
Blog: blog.langchain.dev/agent-protoc...
1 Billion nested loop iterations
Devs be like:
JS devs: "Actually our language is pretty fast"
Python devs: "Come defend our language!"
Go devs: "Needs more optimizations"
C devs: "Just enable the optimizer and watch it gooooo.." 🚀
Devs be like:
JS devs: "Actually our language is pretty fast"
Python devs: "Come defend our language!"
Go devs: "Needs more optimizations"
C devs: "Just enable the optimizer and watch it gooooo.." 🚀
November 19, 2024 at 9:59 AM
1 Billion nested loop iterations
Devs be like:
JS devs: "Actually our language is pretty fast"
Python devs: "Come defend our language!"
Go devs: "Needs more optimizations"
C devs: "Just enable the optimizer and watch it gooooo.." 🚀
Devs be like:
JS devs: "Actually our language is pretty fast"
Python devs: "Come defend our language!"
Go devs: "Needs more optimizations"
C devs: "Just enable the optimizer and watch it gooooo.." 🚀