Technology
Anthropic changes AI training method after Claude Opus 4 incident
|
Anthropic said that training methods and internet data can cause AI models to develop dangerous deviant behaviors.
UK tests first automated landing platform for UAVs at sea
|
He tested an AI platform to help unmanned aerial vehicles automatically take off and land on ships moving offshore.
When natural disasters are covered by fake information created by AI
|
Unusual thunderstorms and hailstorms that have appeared recently are becoming a "land" for fake news and AI-baited videos to spread widely.
Forging subscriber authentication after deploying one-way SIM lock
|
After many subscribers were blocked one-way, forms of impersonating network operators requesting data updates also appeared more and more.
After a two-way SIM is locked, how long can the subscriber be recalled?
|
After being one-way SIM locked, many users are starting to worry about the risk that subscribers may be two-way locked and recovered if the information is not verified.
Information about the transparent back of iPhone 17 Pro leaked
|
Apple is expected to launch the iPhone 17 series on September 9, and the final rumors before G hour continue to appear.
Threads tests long- text posting feature
|
Threads is testing a feature that allows users to attach a long text document to a post.
Reddit reveals a series of smart travel tips to make trips no longer troublesome
|
Smart travel tips from Reddit help your trip become neat, cost-effective, reduce troubles and enjoy your vacation.
AI is exploited as a weapon for cyber attacks, easily bypassing traditional security systems
|
sufferers take advantage of generative AI to develop ransomware that constantly transforms, making it difficult for traditional security systems to detect and prevent in time.
TikTok adds the feature of sending voice messages, photos and videos in chats
|
TikTok is trying to expand its role, not just a place for users to Google through short videos.
Facebook patches exploited vulnerabilities in sophisticated spying campaign
|
Facebook has just patched a serious security vulnerability on iOS and Mac, which was exploited in a sophisticated Escape Campaign targeting nearly 200 users.
Meta considers using both Google and OpenAI models to increase Meta AI power
|
Meta seems to be more interested in hunting for personnel from Google and OpenAI in the artificial intelligence race.
Information leaked about Huawei's rotary screen laptop ahead of IFA 2025
|
Lenovo attracted attention with the Project Pivo laptop prototype with a physical rotation screen between flexible vertical and vertical modes.
personalized brain transplant opens up new hope for chronic pain patients
|
personalized brain transplant detects pain signals and stimulates the brain deeply in real time, opening up new hope for patients with chronic pain.
Software problems cause many Verizon customers in the US to lose connection
|
On August 30 (US time), many users in the US reported Verizon's loss of mobile network connection.
Revealing the group of hackers behind the data leak affecting Gmail
|
The Salesforce data leak puts Gmail at risk, users need to change their passwords, activate two-factor authentication, and be vigilant against scams.
Taco Bell considers using AI at fast food stalls
|
Taco Bell's Director of Technology and Digital Engineering said that the company had a discussion about which stages should and should not be used for AI.
Instagram deploys thu phong management tool for major content creators
|
Instagram has just introduced a new toolkit to help famous content creators easily process the huge number of messages and requests every day.
Meta strengthens AI training, adds measures to protect young users
|
Meta continues to train artificial intelligence tools and deploy additional measures to protect minors.
Google promotes AI startups in Vietnam
|
Through the AI Solutions Lab program, Google has promoted Vietnamese startups.
