BTC $84,635.11 -1.67%
ETH $2,677.31 -1.46%
BNB $766.92 -1.35%
XRP $1.49 -2.09%
SOL $119.54 -2.07%
TRX $0.3348 +0.19%
DOGE $0.0930 -2.82%
ADA $0.2465 -2.61%
BCH $309.61 -1.16%
LINK $14.04 -2.93%
HYPE $88.52 -1.50%
AAVE $179.53 -0.82%
SUI $1.14 -2.35%
XLM $0.2151 -3.17%
ZEC $1,321.60 -4.59%
AAPL $333.32 +0.58%
AMZN $251.73 +1.19%
GOOGL $343.77 +1.14%
MSFT $517.66 +0.25%
META $726.94 -0.20%
NVDA $234.25 +0.96%
TSLA $370.81 +4.18%
SNDK $1,718.31 -3.71%
INTC $117.98 -2.20%
SPCX $158.88 +6.51%
MU $1,067.29 -2.21%
AMD $633.22 +2.27%
BTC $84,635.11 -1.67%
ETH $2,677.31 -1.46%
BNB $766.92 -1.35%
XRP $1.49 -2.09%
SOL $119.54 -2.07%
TRX $0.3348 +0.19%
DOGE $0.0930 -2.82%
ADA $0.2465 -2.61%
BCH $309.61 -1.16%
LINK $14.04 -2.93%
HYPE $88.52 -1.50%
AAVE $179.53 -0.82%
SUI $1.14 -2.35%
XLM $0.2151 -3.17%
ZEC $1,321.60 -4.59%
AAPL $333.32 +0.58%
AMZN $251.73 +1.19%
GOOGL $343.77 +1.14%
MSFT $517.66 +0.25%
META $726.94 -0.20%
NVDA $234.25 +0.96%
TSLA $370.81 +4.18%
SNDK $1,718.31 -3.71%
INTC $117.98 -2.20%
SPCX $158.88 +6.51%
MU $1,067.29 -2.21%
AMD $633.22 +2.27%

mist

All
Article
Flash

first_img Tavus claims that the Griffin model causes 48% of video call participants to mistakenly believe they are interacting with a real person

AI startup Tavus released a new model called Griffin on October 1, referring to it as the first "human interaction model" that can understand and generate face-to-face conversations, paying attention not only to wording but also to expressions and pauses. Tavus stated that in tests, after a one-minute video call with Griffin-Lite, 26 out of 54 participants (48%) believed the other party was a real person; its previous generation system only convinced 1 out of 41 participants.Participants were told they would have a one-minute video call with another person to discuss things they were looking forward to this year, and only at the end of the call were they asked if they suspected the other party was not a real person. Tavus noted that those who became suspicious typically noticed within 20 seconds. This result comes from Tavus's own research page, where participants were recruited through a channel they call an independent research platform. Community comments on X pointed out that these results have not been independently verified and do not meet standard protocols.In the NVIDIA VideoFDB benchmark, Tavus claimed Griffin-Lite ranked first, with a generation dimension score of 3.83, the next best system scoring 2.8, and human reference scoring 3.92; the perception dimension score was 3.73, with the strongest baseline at 3.44 and human reference at 4.2. Tavus stated that NVIDIA independently conducted this evaluation. Griffin supports full duplex, allowing it to listen, see, and speak simultaneously, with an average audio-video delay of 0.43 seconds on the NVIDIA H100 chip. Tavus completed a $40 million Series B funding round led by CRV in November 2025.

SlowMist: FomoPeek versions 1.1–1.2 contain malicious code, which may lead to the leakage of private keys and mnemonic phrases

SlowMist released a security warning stating that it has recently received multiple reports of FomoPeek users' assets being stolen. After a joint investigation with the OKX security team, it was found that some affected users had previously installed or used FomoPeek versions 1.1 to 1.2, which contained malicious code. SlowMist stated that there are modules in FomoPeek unrelated to normal business, one of which includes a kernel exploit framework targeting the iOS system, supporting eight different attack methods that can automatically select the exploitation method based on device model and iOS version. Affected systems include iOS 12 to 18.7 and iOS 26 to 26.1. If the exploitation is successful, the application may break through the iOS sandbox and access and decrypt Keychain data, leading to the leakage of private keys, mnemonic phrases, login credentials, and other sensitive files. In addition, FomoPeek also connects to hidden servers unrelated to its public services and can receive remote commands. SlowMist indicated that its analysis of captured plaintext traffic shows that the related attack functions are currently enabled and will run automatically on a regular basis. SlowMist recommends that users who have installed or used FomoPeek versions 1.1 to 1.2 immediately check for any anomalies in their assets, generate new private keys and mnemonic phrases on trusted devices that have never installed the application, and transfer assets to new accounts as soon as possible, while also upgrading to the latest iOS version and not continuing to use or reinstall FomoPeek.
app_icon
ChainCatcher Building the Web3 world with innovations.