Analysis of tweet claiming industrial-scale distillation attacks by DeepSeek, Moonshot AI and MiniMax. Public reaction: 34.03% supportive, 42.89% confronting.
We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.
Real-time analysis of public opinion and engagement
What the community is saying — both sides
and “stealing” — several replies use jokes and mockery but the tone is sharply accusatory toward the alleged culprits.
, framing it as an AI arms race and national security problem; commenters worry foreign labs could strip safeguards and feed capabilities into military and surveillance systems.
, emergency powers, and coordinated responses among industry and policymakers appear throughout, alongside praise for Anthropic’s decision to publicly name alleged offenders.
and 16 million exchanges, describe distinctive “fingerprints” of distillation at scale, and suggest stricter API controls, rate limits, and account verification as defenses.
survive? That worry about removed guardrails recurs across comments.
systems, improved red‑teaming and vulnerability scanning (e.g., “Claude Code”), and identity layers that can resist Sybil attacks — but many users doubt how easily industrial scraping can be stopped.
some celebrate Anthropic’s bold callout, others warn that the fight will be messy and require sustained investment in security, attribution, and industry norms.
calling out that the company itself trained on massive scraped datasets while now denouncing others for distilling its outputs, and framing the company’s complaint as a double standard.
was processed, the lawful basis under Art.6, whether EU users were informed (Arts.13/
and retention rules (Art.5), and if metadata is used for user-level attribution/tracking.
A large segment champions open-source distillation as legitimate competition and innovation, praising projects like DeepSeek and arguing cheaper, reproduced models are a market correction rather than an “attack.”
Tone across many replies is caustic and mocking — memes, insults, “cry more” refrains and ridicule dominate, signaling strong public scorn and derision toward the announcement.
complaints about pricing, API token costs, the 24k-account incident, calls for KYC or better safeguards, and questions about whether accounts were paying customers or fraudulent.
Creators and rights advocates demand accountability and compensation, with numerous reminders that copyrighted books and individual works were allegedly scraped for training and calls for remedies or legal recourse.
Most popular replies, ranked by engagement
Distillation can be legitimate: AI labs use it to create smaller, cheaper models for their customers. But foreign labs that illicitly distill American models can remove safeguards, feeding model capabilities into their own military, intelligence, and surveillance systems.
it’s only Claude if it’s distilled in the Silicon Valley region of California 😤
@AnthropicAI https://t.co/TdLnwKTaGw
didn't y'all steal the entire internet to train?
These attacks are growing in intensity and sophistication. Addressing them will require rapid, coordinated action among industry players, policymakers, and the broader AI community. Read more: https://t.co/4SVm8K3qou
you trained on the open internet and then call it “distillation attacks” when others learn from you labs that like to preach “open research” suddenly crying about open access this is what happens when intelligence sits behind a centralized api with subsidized tokens
Found something wrong with this article? Let us know and we'll look into it.