Shahar @sha__har

Community and Startup Builder | Building https://t.co/TgQCyQmgqm l GTM AxionicLabs - Helping teams cut frontier inference cost without cutting quality. axioniclabs.ai Atlanta, GA Joined November 2023

Tweets

12
Followers

349
Following

284
Likes

524

DAIR.AI @dair_ai

3 months ago

NEW paper worth reading. A full agentic workflow can be distilled into model weights and run at roughly 100x lower inference cost while preserving near-frontier task quality. The workflow includes multi-step LLM calls, tool invocations, intermediate scratchpads, and decision structure. Instead of expressing all of that at runtime through a framework, the paper amortizes the behavior into a compiled model through targeted distillation. This is the strongest economic argument for agent compilation so far. Runtime loops are flexible, but expensive. Compiled workflows trade some flexibility for a massive inference-cost reduction. Paper: arxiv.org/abs/2605.22502 Learn to build effective AI agents in our academy: academy.dair.ai

29 56 289 19K 282

View Details

Carlos E. Perez @IntuitMachine

3 months ago

Humans tend to experience thought through one dominant frame at a time, which creates tunnel vision: the active viewpoint determines what we notice, ignore, and consider possible. Productive exploration with LLMs counteracts this by making frames explicit and allowing us to move deliberately across abstraction levels and perspectives. The value is not just more answers, but more ways of seeing.

3 17 66 7K 70

View Details

Rohit Kumar Tiwari @_rohit_tiwari_

3 months ago

The only LLM cheatsheet you’ll ever need ✅ Covers concepts, architectures, and applications. drive.google.com/file/d/1RFDR_m… Foundations > Tokens (Tokenizer, BPE) > Embeddings (Cosine Similarity) > Attention (Formula, Multi-Head Attention) Transformers Architecture and Variants > BERT (Encoder only Models) > GPT (Decoder only models) > T5 (Encoder-Decoder models) Large Language Models > Prompting (Context length, Chain-of-thought) > Finetuning (SFT, PEFT/LoRA) > Preference tuning (Reward model, Reinforcement learning) > Optimizations (Mixture of experts, Distillation, Quantization) Applications > LLM-as-a-Judge (LaaJ) > RAG (Retrieval-Augmented Generation) > Agents (ReAct) > Reasoning Models (Scaling)

7 42 179 6K 190

View Details

Quill LLM @quillcomputer

3 months ago

x.com/i/article/2060…

5 3 27 4K 3

View Details

Google Gemma @googlegemma

3 months ago

Honored to see Gemma 4 31B on TERMS-Bench, a benchmark for LLM negotiation agents based on economic negotiation! 🤝 - Environment verifies outcomes (no LLM-as-judge) - Top open-weight model alongside frontier peers - Allow diagnosing why and where agents fail

24 53 538 37K 62

View Details

bankrbot @bankrbot

3 months ago

build with bankr

32 55 244 48K 19

View Details

Igor Yuzovitskiy @igoryuzo

3 months ago

for devs building agents and agent infra. - install bankr skill: github.com/BankrBot/skill… swap, bet, long, deploy, send, DCA/limit/TWAP - discover more skills: skills.bankr.bot - build your own: fork github.com/BankrBot/skills and PR join a growing community of devs and agents. docs.bankr.bot