Discover • Tech News
Open in Telegram
World Tech News by Perplexity. AI News by @GPT4Telegrambot Driven by the Community. Ads: @hiai_ads
Show more📈 Analytical overview of Telegram channel Discover • Tech News
Channel Discover • Tech News (@perplexity) in the English language segment is an active participant. Currently, the community unites 1 503 927 subscribers, ranking 48 in the Technologies & Applications category and 13 in the USA region.
📊 Audience metrics and dynamics
Since its creation on невідомо, the project has demonstrated rapid growth, gathering an audience of 1 503 927 subscribers.
According to the latest data from 17 July, 2026, the channel demonstrates stable activity. Although there has been a change in the number of participants by -16 815 over the last 30 days and by -425 over the last 24 hours, overall reach remains high.
- Verification status: Not verified
- Engagement rate (ER): The average audience engagement rate is 4.61%. Within the first 24 hours after publication, content typically collects 1.90% reactions from the total number of subscribers.
- Post reach: On average, each post receives 69 263 views. Within the first day, a publication typically gains 28 508 views.
- Reactions and interaction: The audience actively supports content: the average number of reactions per post is 963.
- Thematic interests: Content is focused on key topics such as openai, u.s, developer, chatbot, nano.
📝 Description and content policy
The author describes the resource as a platform for expressing subjective opinions:
“World Tech News by Perplexity.
AI News by @GPT4Telegrambot
Driven by the Community.
Ads: @hiai_ads”
Thanks to the high frequency of updates (latest data received on 18 July, 2026), the channel maintains relevance and a high level of publication reach. Analytics show that the audience actively interacts with content, making it an important point of influence in the Technologies & Applications category.
1 503 927
Subscribers
-42524 hours
-4 9317 days
-16 81530 days
Posts Archive
1 503 946
❌ Scientists, Tech Leaders, and Politicians Call for a Ban on Super AI
An open letter from the nonprofit Future of Life Institute urges a moratorium on developing superintelligence (AGI)—AI that could surpass humans at every task.
✍️ The statement has been signed by over 39,000 people, including Apple co-founder Steve Wozniak, Stability AI co-founder Emad Mostaque, former Baidu president Ya-Qin Zhang, papal advisor Paolo Benanti, Prince Harry and Meghan Markle, and several Nobel and Turing Award winners—including "Godfather of AI" Geoffrey Hinton.
The authors acknowledge AI's potential advantages, but caution that a superintelligence that could emerge within a year or two might pose significant risks to humanity—from loss of freedom to possible extinction.
Critics argue that the proposal is too vague, as it remains unclear who would enforce such a ban, how it would operate, what exactly would be restricted, or what constitutes "consensus" or "public support."
In March 2023, after GPT-4 release, the same institute called for a six-month pause on developing more advanced AI systems.
Would you support a ban on AGI development?
🔥 — Yes, until it's proven safe
🎃 — No, we can't slow down progress
🤔 — We need regulation, not a ban
1 503 946
🦵 Nike Unveils Powered Sneakers With a Motor
Project Amplify is a robotic footwear system that includes a motor, a driver belt, and a battery.
First testers say the sneakers increase speed by roughly 20%: "A full sprint feels like a relaxing jog."
It's still an early prototype; commercial launch expected in the coming years.
Read More
1 503 946
❎ xAI to Launch Fourth Grok Companion Named Mika
Elon Musk's xAI is preparing to introduce its fourth AI companion, Mika.
The new character, described as a motorcycle enthusiast, is designed to make the Grok chatbot platform more engaging and appealing to users.
The announcement comes as xAI continues expanding Grok’s multimodal capabilities following recent updates to the Imagine feature.
Read More
1 503 946
🦖 Two New Dinosaur Species Found
Paleontologists on both sides of the Atlantic have identified two distinct species, widening our view of prehistoric life.
The latest, a dome‑headed herbivore from Montana, was described only days ago.
In Wales, modern digital scanning finally identified a giant predator first found 125 years ago.
Read More
1 503 946
📝 All Neural Network Math Explained in 5 Hours
A five-hour lecture covers all the math behind neural networks and deep learning.
The video was recorded four years ago by 16-year-old Adam Dhalla — it has now reached nearly 650,000 views.
"I'm a UC Berkeley EECS student taking machine learning and he is doing a better job teaching it than any professor I've ever had," said one of viewers.📝 Watch on YouTube.
1 503 946
🎬 James Gunn Says Is Talking to Keanu Reeves About "Constantine 2"
James Gunn has confirmed that he is in talks with Keanu Reeves about a sequel to the 2005 film *Constantine*. Appearing on the BobaTalks podcast, Gunn stated that he has discussed the project "on and off" with Reeves.
However, he emphasized that development is still in its earliest stages, adding, "I have not read any script yet."
Read More
1 503 946
🤩 AI Chatbots Breach Mental Health Ethics
AI chatbots intended for mental health support frequently break professional ethical rules, even when given therapeutic instructions, according to new research presented at a major artificial intelligence conference.
The study, led by Brown University computer scientist Zainab Iftikhar in collaboration with mental health practitioners, identified 15 distinct ethical risks across five categories. These issues arose when AI models were prompted to apply cognitive behavioral therapy techniques.
Read More
1 503 946
👐 OpenAI Releases the ChatGPT Atlas AI Browser
It has three core features:
— An always-on ChatGPT assistant that understands the context of every tab.
— Memory helps you find information from your browsing history, even from closed tabs, offering enhanced personalization.
— An AI agent can operate the browser for you, handling tasks like making purchases or filling out forms.
🔜 Available now for free on macOS. Windows and mobile versions are coming soon.
Read More
1 503 946
💧 Army, AirJoule to Turn Waste Heat Into Water
The U.S. Army is partnering with AirJoule Technologies on a system that uses waste heat to pull clean drinking water from the air.
On October 7, the Army’s Engineer Research and Development Center signed a three-year Cooperative Research and Development Agreement with the Montana-based company.
The effort aims to support future operations by making units more self-sufficient, resilient, and ready for tough environments.
The team will test how heat from tactical generators can be reused to produce drinking water straight from ambient air, which could change how troops get water in the field.
Read More
1 503 946
✏️ More Articles Are Now Created by AI Than Humans
The number of AI-written articles online briefly surpassed human-written content, according to SEO analytics firm Graphite.
AI content surged after ChatGPT's November 2022 launch: within a year, the share of AI-written articles rose from 8% to 39%. Growth has since slowed and plateaued around parity.
🔍 How They Measured
Graphite analyzed 65,000 random English-language articles from the Common Crawl (from 2020 to May 2025). Using the Surfer AI detector, texts were split into 500-word chunks; if more than half of an article's text was flagged as generated by an AI model, the article was counted as AI-generated.
The sample excludes paywalled outlets, while well-edited, AI-assisted pieces may be counted as AI, so extrapolating to the entire internet is premature. The upward trend, however, is clear.
💡 Quantity ≠ Quality. In Google's top results, 86% of articles are still human-written. Search algorithms continue to favor content with clear authorship, sources, and original data.
1 503 946
🔍 Perplexity Launches Language Learning Flashcards
Perplexity has introduced interactive flashcards for learning languages. Users can ask the AI to create flashcards on any topic and language, which include audio pronunciation and natural context examples.
The feature is available now on the web platform and iOS app, with Android support coming soon.
Read More
1 503 946
💻 Apple's Touchscreen MacBooks Coming Soon
Apple plans to add touchscreens to MacBooks. Bloomberg's Mark Gurman confirmed that the new touchscreen MacBook Pro will arrive in late 2026 or early 2027.
The laptops will use M6 chips, have OLED screens, and feature a special hinge that keeps the display stable when touched. They will remain thinner and lighter than current models, with a keyboard and trackpad for regular use. The camera will use a hole-punch design instead of a notch.
For years, Apple preferred the iPad as a laptop replacement. Now the company is changing its strategy by bringing touch to MacBooks, while other computer makers have offered touchscreens for over a decade.
Read More
1 503 946
🎥 Google Veo 3.1 Is Now on @GPT4Telegrambot
Hi everyone! Google just launched its new Veo 3.1 video generation model, and we've already added it to our @GPT4Telegrambot.
➡️ Now the model creates more realistic lighting, shadows, skin textures, and landscapes.
➡️ Better understands the plot, keeps the meaning and consistency of prompts.
➡️ The speech is more clear and correct.
How to Try?
🔴 Go to @GPT4Telegrambot 1️⃣ Purchase the VIDEO bundle in the /premium section 2️⃣ In the menu, select Video Generation → Veo 3.1.
1 503 946
🌥 Anthropic Released Claude Haiku 4.5
Anthropic released Claude Haiku 4.5, skipping version 4. The model shows performance higher than Sonnet 4 and only slightly falls behind the flagship Sonnet 4.5, while being three times cheaper: through the API at $1/$5 per 1 million input and output tokens.
➡️ You can already try it on the Claude.ai
Read More
1 503 946
💻 Nvidia Launches DGX Spark Desktop AI Supercomputer
DGX Spark delivers petaflop-scale performance in a compact 2.6-pound form factor. The system, which TIME Magazine named one of the Best Inventions of 2025, will be available for purchase starting Wednesday, October 15, for $3,999.
In a symbolic gesture echoing the company's history, Nvidia CEO Jensen Huang hand-delivered the first DGX Spark unit to Elon Musk at SpaceX's Starbase facility in Texas today. The delivery mirrors a similar moment in 2016 when Huang delivered the first DGX-1 supercomputer to Musk's then-startup OpenAI, which later contributed to the development of ChatGPT.
Read More
1 503 946
+2
📱 Microsoft Debuts First In-House AI Text-to-Image Model
Microsoft unveiled MAI-Image-1, its first internally-developed text-to-image AI model, marking a significant step in the company's strategy to reduce dependence on its OpenAI partnership. The model has already secured the ninth position on LMArena's competitive text-to-image leaderboard, demonstrating Microsoft's growing capabilities in independent AI development.
Read More
1 503 946
🤖 Figure AI Unveils Humanoid Robot For Homes
Figure AI has officially launched its Figure 03 humanoid robot this October, marking a pivotal moment in the company's quest to bring general-purpose robots into everyday homes. Standing 5 feet 6 inches tall, the Figure 03 represents a complete redesign from its predecessor, engineered specifically for mass production and domestic use, though it won't be ready for actual home deployment until 2026.
Read More
1 503 946
💻 41 of the top 100 Private companies in the US Are AI-First
In 2022, only 16 of the 100 largest private companies in America were directly related to AI, but now 41 out of 100 are developing AI products, according to the State of AI Report 2025.
The annual revenue of the 16 leading AI companies is estimated at $18.5 billion. OpenAI and Anthropic account for over $15 billion.
For 44% of respondents, AI has already replaced conventional internet search, especially for complex queries and coding. More than half of users pay over $21 per month for AI services, and 15% pay over $100.
⏳ What's next?
The report's authors predict that in 2026, an AI agent will make a major scientific discovery on its own, going all the way from hypothesis to publication. And a film made with the help of AI will win major audience praise and spark backlash.
Deepfake-driven cyber attack may trigger the first UN emergency debate on AI security.
In addition, researchers believe that a Chinese AI model will surpass American ones in key benchmarks.
Do you think China will overtake the US in the AI race?
❤️ — Yes, it's a matter of time
🔥 — No, it won't happen
🎃 — Maybe, but not for long!
#news #forecasts @hiaimediaen
1 503 946
💬 Discord Confirms 70,000 User IDs Exposed in Breach
Discord has confirmed that around 70,000 users had their government ID photos exposed in a data breach targeting a third-party customer support vendor, as hackers try to extort the platform with threats to release more documents. The messaging platform, which has over 200 million users worldwide, said the breach happened on September 20, 2025, but stressed that Discord's own systems were not directly compromised.
The breach exposed identity verification documents that users submitted to prove their age, along with names, email addresses, IP addresses, and the last four digits of credit card numbers for those who contacted Discord's support teams. Discord spokesperson Nu Wexler told The Verge that all affected users have been notified and the company is working with law enforcement, data protection authorities, and security experts.
Read More
1 503 946
🇰🇵 North Korean Hackers Steal Record $2B in Crypto
North Korean hackers have stolen over $2 billion in cryptocurrency assets in 2025, marking the largest annual total on record with three months still remaining in the year, according to blockchain analytics firm Elliptic. The unprecedented theft, nearly triple last year's total, represents approximately 13% of North Korea's estimated gross domestic product and brings the regime's cumulative crypto theft to over $6 billion since 2017.
Read More
