Roblox downloaded 145GB models from us (including adult content) without asking
open3dlab.com ·
Open3DLab, SFMLab and SmutBase say a crawler identifying itself as “trellis2-data-sourcing/1.0” and listing a Roblox employee’s email address downloaded 145.5 GB of 3D assets on October 15 and 16, 2026. The sites say the crawler visited 18,926 project pages in 26,429 requests over 34 hours, then copied 1,027 files from 712 projects. It operated from a single AWS address and retrieved, but did not follow, the sites’ robots.txt rules, according to the operators.
The downloaded files included 364 files totaling 54.8 GB from 224 SmutBase projects, an adult-only site. Including adult-marked SFMLab projects, 345 of the 712 projects and 62.6 GB of the downloaded material were adult content. The operators say 423 of the affected projects carried non-commercial licenses. They blocked the crawler on October 18; it resumed requests on October 19 and 20 but was blocked. The sites say Roblox did not reply to two emailed requests for comment.
The operators believe the collection may have supported Roblox’s 3D foundation-model work, but say they cannot prove the purpose or whether files entered a model. Roblox introduced its Cube generative AI system in March 2025 and began rolling out Cube-based 4D generation in February 2026. The crawler’s stated email belongs to a listed coauthor of Roblox’s Cube paper, according to the sites.
What people are saying
The thread spent less time on the raw 145GB figure than on the norms governing AI data collection.
What readers liked
- Adult 3D assets could plausibly help train filters that catch sexual uploads, turning uncomfortable source material into stronger platform safety…
- Some commenters argue that excluding adult material can weaken generative models' depiction of human bodies, so blanket purges may degrade broader…
Concerns raised
- Ignoring stated crawl restrictions and non-commercial licenses turns a community asset library into uncompensated corporate input, precisely the…
- Treating public download links as blanket consent erases the distinction between sharing work with a community and feeding it into a commercial…
Read full discussion analysis
Source: Hacker News · Based on 15 sampled comments · Analyzed
Discussion overview
The thread spent less time on the raw 145GB figure than on the norms governing AI data collection. A sizable group rejected the idea that publicly downloadable files are automatically fair game: the site says the crawler retrieved but disregarded robots.txt, site rules, and many non-commercial creator licenses. For them, industrial-scale collection is not equivalent to an ordinary user downloading a model; it converts a community archive into unpaid corporate training or product input while potentially making such archives impossible to host openly.
Others treated the complaint as overstated or internally awkward. They argued that a site hosting fan-made and sometimes questionably licensed assets has weak standing to invoke moral ownership, and that a modest public download is neither technically remarkable nor necessarily evidence of model training. Several readers also questioned the emphasis on adult assets, suggesting Roblox may have collected them to build moderation systems that identify or reject sexual content. A minority added that adult examples can improve generative models' understanding of human anatomy even when the intended output is non-explicit.
The sharper criticism was directed at the broader AI-industry premise: firms present indiscriminate ingestion of others' work as progress toward AGI, while critics see a commercial extraction project with neither consent nor payment. Practical anti-scraping proposals drew skepticism because cloud blocking can exclude legitimate users and determined collectors already use residential proxies.
Points of disagreement
The central dispute is whether openly accessible community assets may be bulk-collected for an AI-related corporate purpose despite explicit restrictions. Defenders of the scrape say availability invites downloading and adult files could support safety moderation; critics say robots rules, non-commercial licenses, and the scale of automated extraction make that argument a license to strip creators of meaningful control.
What readers liked
- Adult 3D assets could plausibly help train filters that catch sexual uploads, turning uncomfortable source material into stronger platform safety controls.
- Some commenters argue that excluding adult material can weaken generative models' depiction of human bodies, so blanket purges may degrade broader visual quality.
- The episode has readers thinking concretely about crawler defenses and the operational limits of relying on public hosting for valuable AI-adjacent datasets.
Concerns raised
- Ignoring stated crawl restrictions and non-commercial licenses turns a community asset library into uncompensated corporate input, precisely the extraction pattern that has soured people on AI.
- Treating public download links as blanket consent erases the distinction between sharing work with a community and feeding it into a commercial model or pipeline.
- The AGI justification drew open contempt: commenters see a small set of companies monetizing everyone else's data, not a public bargain people meaningfully accepted.
- Bulk scraping can make open creator repositories financially and operationally fragile, pushing communities toward paywalls or shutdowns rather than sharing reusable 3D work.
Sources and details
open3dlab.com · Published
Roblox downloaded 145GB models from us (including adult content) without asking
Read report →
Community analysis available for 1 of 2 linked discussions.
Additional discussion searches are still incomplete.
Coverage details and dates
Latest available community analysis:
Summaries: 1 current · 0 awaiting update · 0 not verified
Latest summary update:
Discussion analysis: 1 unavailable
Hacker News · open3dlab.com