Adobe Firefly

129 comments total (top 25 shown, all HackerNews). Direct Firefly discussion is sparse — the dominant thread is a philosophical/legal debate about AI image generation and copyright. The most direct Firefly comment (+908) attacks Adobe's beta messaging: the line 'Once Firefly is out of its beta stage, creators will have the option to use content generated in Firefly commercially' is read as Adobe implying it owns the outputs, when 'under current copyright interpretation, Adobe doesn't own those images... that's not Adobe's permission to give.' Other highly upvoted themes: artists report their work was plagiarized long before AI ('ever since I've been publishing my own illustrations I've always ran an image search... almost every time I've found multiple stolen copies'); a long exchange over whether AI learning an artist's 'style' constitutes infringement ('Nobody owns a style... it belongs to the art critic more than the artist' vs. arguments that style is core expression); consensus that AI-output copyright is legally unsettled ('That's not yet settled'), with the Copyright Office's commission analogy cited to argue prompts alone confer no authorship; and artists advising peers that 'it's naive to expect a reward' because theft predates AI. Technical comparisons in the thread center on DALL-E vs Midjourney vs Stable Diffusion vs Kandinsky 2.1 — commenters stress cross-generator prompt tests are flawed because 'each system benefits from a different approach to prompting,' and one speculates OpenAI keeps DALL-E alive only to stay 'on the list.' Net: user sentiment toward Adobe specifically is distrustful on licensing/IP framing; there are virtually no first-hand reports of Firefly's output quality in this layer.

No aggregate ratings were found for this product during the last harvest.