Developer questions whether confidential-computing AI inference hardware flaws are underestimated
Understand this story
A developer says services claiming to offer private AI inference in Intel SGX/TDX confidential-computing environments may have their security overstated. The RAM interposition demonstrated by tee.fail is too conspicuous in a proper data center, but an attacker only needs to extract a key once via a physical attack, after which they can impersonate that CPU in software. The developer argues that users currently cannot verify whether a TDX attestation quote was forged with such an extracted key.
AI overview based on source reports · Updated 2 days ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AITell HN: Hardware flaws to confidential AI are probably worse than they let on
A Hacker News post questions how users of private AI inference running in confidential computing environments like Intel SGX/TDX can verify that TDX attestation quotes aren't forged using a key extracted from a processor via a physical attack, as described in tee.fail. The author notes limited expertise on the topic and links to the tee.fail section on recovering attestation keys from Intel machines.
Activity for this story
- Current comparable activity
- 3
- Peak comparable activity
- 10Nov 13 04:00
- Change in the past 24 hours
- -50%
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.