Prompt Engineering

AI Topics Discussed on 23 Feb, 2026

—

by

Fraser

AI Topics Discussed on 23 Feb, 2026

Creative & Visual Media

Javi Lopez highlighted the stunning realism of Seedance 6.9, a generative video model capable of producing Oscar-worthy films from a single prompt, dismissing criticisms of AI-generated content as “slop.”

Heather Cooper showcased Kling 3.0’s video generation capabilities in a demo clip.

She also praised node-based workflows for simplifying content creation pipelines, noting she’s building a personal catalogue of them.

Alex Volkov mentioned using Descript for video editing as an alternative to traditional editors.

Machina (@EXM7777) announced upcoming newsletters covering AI content creation for cheap lead acquisition and distribution, positioning it as one of the only AI verticals that will matter in 2026.

Software Development

@levelsio described using Claude Code, an AI coding agent, to complete and deploy a weekly summary feature for nomads.com. The tool summarized top Telegram channels for nomad communities, prioritized by the user’s current location (e.g., Brazil channel for him), and included deep links to chats and profiles.

He noted a few iterations were needed, but an error caused it to send prematurely to 500 members, motivating a quick launch.

Automation & Orchestration

Heather Cooper demonstrated how intuitive node-based systems make building complex workflows, sharing a video example and her ongoing catalogue project.

Machina (@EXM7777) highlighted “openclaw & the future of AI agents” as a key AI vertical for 2026 in upcoming newsletters.

Strategy & Ecosystem

Cristóbal Valenzuela reiterated the profound social and cultural impacts of generative video models, predicting widespread discussion on pixel manipulation technologies.

Alex Volkov reported on Anthropic publicly accusing DeepSeek, Moonshot, and MiniMax of large-scale distillation attacks on Claude using thousands of fake accounts, signaling escalating industry tensions, and speculated that a new DeepSeek model could outperform rivals.

fofr compared Theo’s Quipslop game—where LLMs compete to generate humor—to a free-form Cards Against Humanity, showcasing experimental model applications.

Volkov celebrated WeaveHacks 3 winners, including a high schooler, highlighting AI hackathons for emerging talent, and marked the third anniversary of the Latent Space AI podcast.

Ethan Mollick (@emollick) emphasized collecting hard problems and good ideas now, as AI enables agency without direction, warning against undirected AI use.

Machina (@EXM7777) argued that “skills” are essentially advanced prompt engineering, repackaged prompts that determine quality, dismissing claims that prompt engineering is dead and stressing it as a life-changing skill.

Justine Moore (@venturetwins) shared Anthropic’s announcement of industrial-scale distillation attacks by DeepSeek, Moonshot AI, and MiniMax, involving 24,000 fraudulent accounts and 16 million Claude interactions to extract capabilities.

Ethan Mollick (@emollick) shared a benchmark visualization app showing saturation across metrics, suggesting a “hard takeoff” at least in software.

Ethan Mollick (@emollick) commented on an influential article expected to move AI markets, agreeing with half and disagreeing with the other.