Prompt Engineering

AI Topics Discussed on 09 Feb, 2026

—

by

Fraser

Creative & Visual Media

gokayfem shared an archival video from 2022 highlighting the distinct aesthetic of early Stable Diffusion generations, including quirky artifacts like random blue dots.

Heather Cooper (@hbcoop_) posted AI-generated visuals, including a motivational image captioned “Do it for you” and a video of a character declaring “In time, I will slay more than slime,” showcasing advancements in generative content creation.

fal launched real-time image-to-image editing on their platform, achieving 10+ FPS with FLUX.2 Klein for low-latency, production-ready results.

Justine Moore from a16z shared examples of Seedance 2, a new Chinese AI video model gaining traction on social media, comparing its impact to the birth of photography and noting it passed the Turing Test for video with a prompt generating a photorealistic Olympics gymnastics scene.

Software Development

Guillermo Rauch (@rauchg) highlighted GPT 5.3 Codex (xhigh) achieving 90% on Next.js evaluations, outperforming competitors.

Alex Volkov (@altryne) compared Opus 4.6 and 5.3 Codex, likening them to senior engineers—one hyper-technical and delivery-focused, the other collaborative and empowering.

Volkov also noted exploding usage of Claude Code, with SemiAnalysis reporting daily spend hitting $6k on Superbowl Sunday, up from under $1k two weeks prior.

Dan Shipper shared Every’s definitive guide to “compound engineering,” aimed at learning engineering in the AI age.

levelsio noted AI now handles all his coding, shifting accomplishments toward management and prompting feelings of unaccomplishment.

Omar Sar discussed FullStack-Agent, a multi-agent system for end-to-end full-stack web development with innovations like development-oriented testing, achieving strong benchmark results on backend and database tasks.

OpenRouterAI introduced Aurora Alpha, a stealth model optimized for fast reasoning in coding assistants and real-time conversational apps.

Automation & Orchestration

Alex Volkov (@altryne) questioned the productivity of OpenClaw users, asking what percentage of chats involves debugging bots versus actual work, echoing concerns about over-optimization on the tool itself rather than building products.

Dan Shipper highlighted OpenClaw usage at Every, including three agents discussing and pitching stories in Discord with wild results, a dedicated claws channel, and interest in a battle royale stream.

goose_oss shared a custom OpenClaw alternative using goose as backend on Raspberry Pi via RPI method to prevent API hallucinations.

Omar Sar showcased Claude Code building a 10K-line agentic video editing app locally using Claude Agent SDK.

Riley Brown announced Claude Code integration in vibecodeapp, enabling single-prompt creation of fully agentic applications without an API key, demonstrated with Twitter tools and motion graphics.

Machina showcased OpenClaw running 10 Opus-powered agents to autonomously organize a downloads folder, building their own skills and infrastructure.

The Boring Marketer launched an AI tool for generating optimized daily ad creatives that learns brand style, competitors, and preferences on autopilot.

Strategy & Ecosystem

Guillermo Rauch (@rauchg) offered to cover Vercel’s bill for Jmail, a high-traffic app (609th on Vercel) archiving Epstein emails, praising its rapid build and global impact while proposing optimizations for massive scale.

Volkov marveled at “vibecoding” enabling feats like Jmail’s high-quality archive, questioning if such projects were feasible five years ago at this speed and cost.

Simon Willison linked HBR research on AI intensifying work, leading to burnout by accelerating tasks without reducing load, a phenomenon he experiences.

Jonathan Fischoff commented on Baker McKenzie’s ~700 layoffs (~10% staff) due to AI.

Ethan Mollick discussed LLM limitations in crafting satisfying medium-length stories compared to proofs, code, or reports, and argued “judgement” won’t remain a human-AI divide; he also emphasized leadership challenges in AI adoption to avoid burnout and reorganize work.

Machina shared a ChatGPT usage chart from an OpenAI paper, asking for comparisons to Claude and Gemini.