Forwarded from Tech Star Updates | Leaks & TechNews (Tech Star)
Poco F7 Pro (24117RK2CG) certified by Thailand's NBTC. Looks like its heading for an early launch.
Poco F6 Pro was NBTC certified in March 2024 and debuted in May.
Probably, F7 Pro likely to debut to by late March or April of this year.
#Poco #PocoF7Pro #PocoF7series #PocoF7 #PocoF7Ultra
Poco F6 Pro was NBTC certified in March 2024 and debuted in May.
Probably, F7 Pro likely to debut to by late March or April of this year.
Only for global not for india
#Poco #PocoF7Pro #PocoF7series #PocoF7 #PocoF7Ultra
👍3
Live at 12 AM
Vivo T3 Pro 5G (8 GB Ram + 128 GB Storage) at ₹19,999 (effective)
https://fkrt.cc/EUH66M
₹3000 Off With HDFC/SBI Cards/Fk Axis Credit Cards + 5% Extra Cashback
Extra Off Using No Cost Emi Transaction
Vivo T3 Pro 5G (8 GB Ram + 128 GB Storage) at ₹19,999 (effective)
https://fkrt.cc/EUH66M
₹3000 Off With HDFC/SBI Cards/Fk Axis Credit Cards + 5% Extra Cashback
Extra Off Using No Cost Emi Transaction
👍3
🚨 NVIDIA has now lost a record $600 billion in market value today, the largest single-day loss in U.S. history. This drop is reportedly due to competition from Chinese AI startup DeepSeek, which released its advanced AI in the U.S. market for free.
ChatGPT ai running on NVIDIA's chips.
#DeepSeek #Nvidia #ChatGPT
ChatGPT ai running on NVIDIA's chips.
#DeepSeek #Nvidia #ChatGPT
😱10👍6😢1
The Google Play Store is going to show a "Verified" badge for some VPN apps.
These badges will appear on store listings for VPN apps that "prioritize user privacy and safety, help users make more informed choices about the VPN apps they use, and build confidence in the apps they ultimately download."
VPN apps have to earn the badge by "adhering to the Play safety and security guidelines" and successfully completing a "Mobile Application Security Assessment (MASA) Level 2 validation." They also have to have at least 10k installs and 250 reviews, be published for at least 90 days, and submit a Data Safety section declaration while opting into independent security review and encryption in transit.
Nord, hide dot me, and Aloha Browser are among the few to secure the badge. You can easily find more by searching "VPN" in the Play Store.
Out of curiosity: What VPN are you currently using, and why did you choose it?
These badges will appear on store listings for VPN apps that "prioritize user privacy and safety, help users make more informed choices about the VPN apps they use, and build confidence in the apps they ultimately download."
VPN apps have to earn the badge by "adhering to the Play safety and security guidelines" and successfully completing a "Mobile Application Security Assessment (MASA) Level 2 validation." They also have to have at least 10k installs and 250 reviews, be published for at least 90 days, and submit a Data Safety section declaration while opting into independent security review and encryption in transit.
Nord, hide dot me, and Aloha Browser are among the few to secure the badge. You can easily find more by searching "VPN" in the Play Store.
Out of curiosity: What VPN are you currently using, and why did you choose it?
👍5
Xiaomi HyperOS 2.1 global first batch of devices confirmed
👍1
Google play system update available
Updated to January 2025 patch
Size - 87 MB
Updated to January 2025 patch
Size - 87 MB
👍8
Android 16 improves how hearing aids deal with noisy environments
Android 16 brings new features for LE Audio hearing aids like:
- Bluetooth ambient volume controls
- Ability to switch which mic is used for voice calls
Android 16 brings new features for LE Audio hearing aids like:
- Bluetooth ambient volume controls
- Ability to switch which mic is used for voice calls
👍4
MiCTS_2.2.apk
2.3 MB
☄️ New Version Of 🌐 Google CTS Is Available For Public Release
🔗 What's New
1. Bugs Fixes
2. Improved App Stability
3. Changed App Logo And Name To MiCTS
4. New Version Number To 2.2
⚙️ Note: If Guy's Yu Is Using Previous Version Pls Uninstall That One Then Install This Version...
🔗 What's New
1. Bugs Fixes
2. Improved App Stability
3. Changed App Logo And Name To MiCTS
4. New Version Number To 2.2
⚙️ Note: If Guy's Yu Is Using Previous Version Pls Uninstall That One Then Install This Version...
⚡1
ANOTHER SCAM BY JIO!
If you buy the ₹448 voice-only plan and think you can add the ₹69 data booster plan , you can’t. You need an active data plan to use booster plans. There are data-only plans, but none with 84 days validity. Even a 28-day plan will cost you ₹175 extra.
Not every1 needed 1.5 GB data/D.
If you buy the ₹448 voice-only plan and think you can add the ₹69 data booster plan , you can’t. You need an active data plan to use booster plans. There are data-only plans, but none with 84 days validity. Even a 28-day plan will cost you ₹175 extra.
Not every1 needed 1.5 GB data/D.
👍7🗿4
Starlink has reportedly submitted its formal acceptance of the key conditions set by the Indian government in order to secure a license for launching satellite broadband services in the country
More specifically, the company has accepted the security and data storage requirements mandated by the government (to store all the user data within the country)
#Starlink #India
More specifically, the company has accepted the security and data storage requirements mandated by the government (to store all the user data within the country)
#Starlink #India
👍3
This media is not supported in your browser
VIEW IN TELEGRAM
Thanks to advancements in technology you can now watch advertisements while fueling up your car.
🤣6👍1
DeepSeek-FAQ
2. Are the models of frontier companies always advantageous?
(Current situation for development companies facing urgent challenges)
Q: So, does that mean V3 is a frontier model (at the base level)?
A: It certainly seems to be on par with 4o and Sonnet-3.5, and it’s definitely far above LLaMA.
However, there’s a high probability that DeepSeek distilled models like 4o and Sonnet to create training tokens.
Q: What is distillation?
A: Distillation is a method of extracting the understanding of another model. You feed various inputs into the “teacher” model, then use its outputs to train the “student” model. Most research labs explicitly forbid this kind of distillation. However, it’s very common—not just at DeepSeek, but at numerous other places as well. That’s why new models on the level of 4o and Sonnet keep appearing. Frankly, it’s so commonly used that it would be surprising if no one did it.
Q: Then aren’t top-ranked models at a disadvantage?
A: Exactly. Leading research labs can’t use this method to expand the frontier because they’re bound by their own restrictions, although they might use it for optimizing their own models. The downside is that if they do allow distillation in any form, other labs can free-ride continuously. The cost of developing cutting-edge models is borne solely by these frontier labs.
As a result, even if huge sums are invested in creating leading-edge models, they can be quickly copied through distillation, making it difficult to recoup those investments. The models are commercialized and become widespread almost immediately.
It seems that this is the very reason Microsoft and OpenAI appear to be moving toward a split. Even if you spend 100 billion dollars developing the latest model, it quickly depreciates once it’s copied, making it nearly impossible to recover the costs.
(Be sure to read the entire text. It’s truly insightful.)
2. Are the models of frontier companies always advantageous?
(Current situation for development companies facing urgent challenges)
Q: So, does that mean V3 is a frontier model (at the base level)?
A: It certainly seems to be on par with 4o and Sonnet-3.5, and it’s definitely far above LLaMA.
However, there’s a high probability that DeepSeek distilled models like 4o and Sonnet to create training tokens.
Q: What is distillation?
A: Distillation is a method of extracting the understanding of another model. You feed various inputs into the “teacher” model, then use its outputs to train the “student” model. Most research labs explicitly forbid this kind of distillation. However, it’s very common—not just at DeepSeek, but at numerous other places as well. That’s why new models on the level of 4o and Sonnet keep appearing. Frankly, it’s so commonly used that it would be surprising if no one did it.
Q: Then aren’t top-ranked models at a disadvantage?
A: Exactly. Leading research labs can’t use this method to expand the frontier because they’re bound by their own restrictions, although they might use it for optimizing their own models. The downside is that if they do allow distillation in any form, other labs can free-ride continuously. The cost of developing cutting-edge models is borne solely by these frontier labs.
As a result, even if huge sums are invested in creating leading-edge models, they can be quickly copied through distillation, making it difficult to recoup those investments. The models are commercialized and become widespread almost immediately.
It seems that this is the very reason Microsoft and OpenAI appear to be moving toward a split. Even if you spend 100 billion dollars developing the latest model, it quickly depreciates once it’s copied, making it nearly impossible to recover the costs.
(Be sure to read the entire text. It’s truly insightful.)
👍1