llm 0.35
Release: llm 0.35 New OpenAI model: gpt-6-astra for GPT-6 Astra. Tags: openai, llm, gpt-6-astra
Daily edition · Frameworks
The daily ledger
TILens turns technical updates into a focused daily brief: official releases, trusted reporting, and practitioner analysis, deduplicated and organized by topic.
18 articles · 7 sources · 4 papers ·
Top topics: AI · JavaScript · React
Creepy crawlies
Creepy crawlies Konstantin Ryabitsev discusses how bad the "background radiation" of abusive crawlers has become from the perspective of git.kernel.org, the official Git repository for the Linux kernel: TL;DR: we spend…
Quoting Jakub Pachocki
The strongest argument I see for continuing to train much smarter models quickly is the need to build defensive systems against the dangers posed by other AI. [...] We will need powerful, aligned AI for defense; to…
Video compressor
Tool: Video compressor I recorded a short demo video of my Equal Earth animation on my phone and wanted to publish an optimized version of that video (using FFMPEG) on my blog, so I had Claude Fable 5.1 in Claude Code…
Mercator ↔ Equal Earth
Tool: Mercator ↔ Equal Earth I got curious about the Equal Earth map projection that was recently voted on at the UN so I had GPT-6 Astra (medium) in ChatGPT Work build me this animated transition between Mercator and…
How to Build an AI Chatbot with Gemini and Vercel Serverless Functions 🚀
A couple of months back, I built a chatbot application using React, Node.js, and Vercel Serverless Functions, which I used in my web app, buildcv.makeadifference.app. In this tutorial, I'll walk you t
Query-Aware Code Compression for LLMs: 1.3M LOC to 1,500 Lines in 2.8s
Query-Aware Code Compression for LLMs: 1.3M LOC to 1,500 Lines in 2.8s The Problem You want to ask an LLM a question about your codebase. You have three options: Paste everything — Repomix dumps your entire repo. For…
How I Made My React Website Load Faster by Fixing Images
The easiest performance win in React isn't always your JavaScript. Sometimes it's your images. I used to think performance optimization meant splitting bundles, memoizing components, and chasing Lighthouse scores. Then…
React 19 useActionState: 3 Silent Trapdoors to Avoid
Introduction React 19 has introduced a suite of powerful hooks designed to simplify form handling and state management. Among these, useActionState is perhaps the most anticipated. It promises to eliminate the tedious…
I saved for 8 months to buy a used laptop to build this: A <5KB alternative to bloated error trackers
Hey everyone 👋 I’m a computer science student, and today I’m putting something into the world that I spent months building alone in my bedroom: SnapTrace. I didn't have money for expensive courses, venture capital, or a…
How to serve dynamic user announcements with static SSR performance
Have you ever loaded a SaaS app and seen an announcement banner flicker in 500ms after the page loads, shifting the entire layout down by 60px? That Layout Shift (CLS) destroys user experience and SEO rankings. Here is…
Building an Automated Background Removal Micro-Service with Next.js and Canvas API
As developers, we often run into user-flow bottlenecks where people need to upload profile pictures, product shots, or signatures, but their images come with messy, non-transparent backgrounds. Building a robust image…
Building KORTEX: How We Engineered a Next.js 16 B2B AI Agency Architecture for $10k/Month Retainers
Selling B2B AI automation services in 2026 requires more than a standard portfolio page. Enterprise clients and mid-market SaaS founders are no longer impressed by static text or generic pitch decks—they demand…
Tuning Collective Patterns to Alleviate Congestion in Shared AI Clusters
arXiv:2609.04417v1 Announce Type: cross Abstract: Distributed AI training involves recurring rounds of data exchange between multiple pairs of GPU nodes. Slowdown in even one flow due to congestion can cause the entire…
MineExplorer: Evaluating Open-World Exploration of MLLM Agents in Minecraft
arXiv:2605.30931v4 Announce Type: replace Abstract: Multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and action generation. However, their ability to sustain exploration…
SiLR: Structure-Preserving Admission and Process Reward for LLM Tool Agents
arXiv:2609.04629v1 Announce Type: cross Abstract: A runtime gate for an LLM tool agent is usually cast as a filter. In a ReAct loop a rejected proposal is followed by another at the same state, so the gate is a search…
Iris: Climbing to the Search Frontier
arXiv:2609.04304v1 Announce Type: new Abstract: We present Iris-mini and Iris-pro, two search agents trained at the 35B-A3B and 397B-A17B scales, together with the data pipeline and training recipe behind them. Tasks…
Bob Belderbos: 5 design patterns used in my new habit tracker app
There are plenty of habit trackers, but I wanted to build mine with constraints, a calendar view and habit streaks. So I built commitgraph: a small Django + HTMX app. Here are five Python design and testing patterns…