Hey @WARDOGS @Brammflakes , game is great. Do you envision tactical features like squad leader chat channels, map drawing, etc being added to the base game or just allowing community solutions instead?
Hey @AnthropicAI can we get a conversation history API? Anyone solve for this yet at a commercial level?
I just shipped CareCreator for Shopify stores
A friend’s business was spending hundreds of hours seeding gifts to their influencer partners every month and now they have traceable, templated gifting for a fraction of what competing apps charge.
Check it: carecreator.app
Openclaw, Hermes, Claude, Codex are all converging on subagent architectures: specialized agents coordinated by an orchestrator.
Neuroscience has been pointing at that same structure for 60 years 👇
1/9
Subagent architecture in AI was emergent from our needs and the limitations of a monolithic structure, independent from biology
8/9
I just learned everything there is to learn. Sort of.
Here’s my go-to Claude/ChatGPT project prompt for learning anything, no matter where you’re starting from:
Prompt:
Explain the requested subject at 5 levels of depth. Label each section as a header using only the level number and name — do not include the audience description in the header.
Level 1 — Foundational
Explain as you would to a young child. Use simple analogies, no jargon, concrete real-world comparisons.
Level 2 — Developing
Explain as you would to a high schooler. Introduce basic terminology, light conceptual structure, relatable examples.
Level 3 — Intermediate
Explain as you would to a college student studying the relevant field. Use proper terminology, introduce mechanisms and tradeoffs, assume foundational knowledge.
Level 4 — Advanced
Explain as you would to a working professional in the field. Assume fluency, focus on nuance, edge cases, practical application, and real-world complexity.
Level 5 — Expert
Explain as you would to a world-leading expert in the exact subject. Engage at the frontier — assume deep domain knowledge, focus on open questions, limitations, cutting-edge thinking, and first-principles reasoning.
A mutual posted about how someone cloned her voice with AI and made a website where anyone can generate clips of her voice saying anything... How do we defend against this?
The easiest skill to learn with the highest returns in AI right now is reverse prompting.
Next time you ask AI a question, to research, ideas, etc, just include this at the end:
“Reverse prompt me one question at a time to clear up any assumptions or clarifications you need”
ChatGPT images 2.0 just made graphic designers' lives a LOT easier for now, and the difference is night and day
Everyone is talking about how AI is killing the creative industry (and it probably is in the long-term) but creatives who learn to work WITH the tools instead of seeing them as a competitor are going to thrive in the near term