AI is bringing more leaks and capabilities to more people, which puts higher security pressure on the Bitcoin perimeter software. In an interview with Decrypt, Bitcoin Red Team member Calle stated that the purpose of the team was to identify the AI auxiliary security risks in Bitcoin ecology as soon as possible before the attacker.
Checking isn't about protocol.
Calle stated that the team had not found any problems with the Bitcoin Agreement itself, and that what was really worrying was the wallets, applications, services and other software around Bitcoin. For most users, day-to-day contact is not with bottom protocols, but with these outer tools, and therefore risks are more easily concentrated here.
Bitcoin Red Team consists of some 20 to 25 volunteers, including several Bitcoin developers and privacy protocol developers. Calle stated that the team would receive both security scanning requests from project participants and take the initiative to screen the targeted items. According to him, the team has scanned itself most of the important items in the open-source environment of Bitcoin.
Coldcard, speed up after the incident.
Calle said the team was formed after the Coldcard cold wallet was used. At that time, RochorWatch, CEO, Rob Hamilton, began to examine a number of bitcoin projects, followed by more developers and researchers.
He also mentioned that, following the release of the enhanced Chinese-language AI model, there had been a marked acceleration of the defensive rhythm in the area of cybersecurity. In the case of Kimi K3, for example, such models have increased the capacity of both the defensive and the attacking parties and have further increased the urgency of security research.
The Chinese model is more used in safety research.
In model selection, Calle indicates that teams use the Chinese AI model much more frequently than the US model in safety studies. The reason for this is not that the former must be stronger as a whole, but that the United States model usually imposes stricter security restrictions, involving the detection of loopholes, the use of pathways or the repair of proposals, which make it easier to refuse to respond.
He claims to have experienced a similar situation before joining the Red Army. Some United States models are not only reluctant to assist in the search for loopholes, but may refuse to assist in the analysis of rehabilitation programmes even after the developers have identified the problem. By contrast, the Chinese model is more available for such tasks and is therefore used more frequently for safety purposes.
AI is lowering the threshold.
Calle believes that AI is weakening the information gap in software security in the past. A simple loophole, which would have required a strong technical background to do so, may now be done independently by people with limited experience, supported by AI, with the entire path being compressed from finding problems to using loopholes.
He also indicated that the encryption industry might have suffered such changes earlier than other industries, as the attackers had direct economic motivation for “Internet money”. That is why the Bitcoin ecology is facing this round of security pressures earlier, and other industries may encounter similar problems in the future.