Discover • Tech News

💼 60% of Today's Jobs in the U.S. Did Not Exist in 1940

These are the findings of the report recently published by Bank of America. The analysts argue that apocalyptic predictions about AI "sit uneasily with both economic theory and the evidence so far."

The labor market is constantly transforming, economists note. For example, agriculture, which employed about 40% of Americans at the beginning of the 20th century, now accounts for only 1% of jobs. E-commerce hasn't killed retail either: in the U.S., about 16 million people still work in this sector—roughly the same as in the 1990s.

"Adaptability is the new job security," the report states.

💡 According to the bank's estimates, the implementation of AI will affect one in four jobs worldwide—a total of about 840 million positions, and in high-income countries—up to one-third. But this doesn't mean people will lose their jobs: only 2.3% of professions could become fully automated.

Will your profession still exist in 50 years?

❤️ — Yes, definitely
🔥 — Probably not...

📱 Apple Could Be Working on 'Spatial iPhone' With Holographic Display

Rumors suggest Samsung is developing a holographic display that could be used in a future "Spatial iPhone."

According to a leak, the technology uses eye-tracking and a nano-structured layer to create 3D effects without glasses. The screen would function as a normal 4K display, only activating the holographic depth for specific content.

Samsung has been researching this for years, and Apple has explored similar concepts for a long time. However, the project is still in early development, and such smartphones likely won't arrive until around 2030s.

🥴 Singapore Introduces School Caning For Bullying

Singapore has introduced new guidelines allowing schools to use caning as a last resort for male students aged nine and older who engage in bullying, including cyberbullying. The punishment, limited to a maximum of three strokes, will only be administered by authorized teachers following approval from a principal.

International organizations, including UNICEF, oppose such corporal punishment, citing risks to the physical and mental health of children.

Do you support these measures?

❤️ — Yes, it is fair
🔥 — No, these are medieval methods!

💬 GPT-5.5 Instant Released for All

GPT-5.5 Instant has been released for everyone—the new model is now the default in ChatGPT even for free users, replacing GPT-5.3 Instant.

The AI has become smarter, provides more concise answers, adapts better to the user, and hallucinates half as often.

➡ You can try it out in ChatGPT. The update is rolling out gradually.

📹 Kling Motion: Transfer Movement from Video to Photo

Now with @GPT4Telegrambot you can upload any reference video—a TikTok dance, Reels gesture, or a viral move—and Kling will map that motion onto your photo.

How to Try:

➡️ In the @GPT4Telegrambot menu, tap Video Generation.
1️⃣ Select Kling Motion.
2️⃣ Upload the movement video, then add your photo.
3️⃣ Choose the Kling version and quality: 720p or 1080p.
4️⃣ Start the generation.

Pricing depends on the version, quality, and length of the original video.

💉 Cameroon Sees 70% Drop In Child Malaria Cases With Full Vaccination

More than two years after Cameroon introduced the malaria vaccine into its routine schedule, health officials report a 70% decrease in cases among young children who receive the full series of doses.

A WHO official noted that when the vaccine is used correctly alongside other tools, it significantly reduces child mortality. Malaria remains a major health threat in the country, causing most deaths among children under five.

🎓 Oral Exams Are Making a Comeback at U.S. Universities

"You can't just coast through an oral exam with the help of AI," says Chris Schaffer, a professor of biomedical engineering at Cornell University. He introduced an oral defense format after written assignments: no laptops, no chatbots, no paper—the student simply has to explain to the instructor what they did and why.

American universities are increasingly returning to in-person knowledge checks. Students submit nearly perfect essays and solutions, but they can't always explain their own work. The issue isn't just whether a student used ChatGPT—it's whether they actually learned anything at all.

ℹ️ For example, NYU Stern School of Business is already testing a voice AI assistant that administers exams to students.

💡 The next stage of education may not be a total ban on AI, but a new way of verifying authorship and understanding: can a student defend the work they turned in?

Are oral exams necessary?

❤️ — Yes, it's a real test
🔥 — No, it's a relic of the past

🖥 OpenAI's Brockman Says AI Now Writes 80% of Code

OpenAI President Greg Brockman told a Sequoia Capital audience this week that agentic coding tools have undergone a dramatic leap, going from writing 20% of code to 80% over the course of December 2025, a shift he described as moving AI from a convenience to a necessity.

🐈 A Digital Cat Stops Your Doomscrolling

A Japanese developer released a free browser extension called Cat Gatekeeper that fights doomscrolling by blocking your social media feed with a virtual cat for five minutes once your time limit is reached.

You can install it here.

🕹 Mom forces her son to slam his PS5 after he repeatedly slammed their kitten on the ground

The boy drops and repeatedly smashes the console on her command after allegedly throwing the family cat multiple times.

She says it's meant to address his anger issues, forcing him to repeat it about eight times before telling him to clean up the pieces and throw them away.

Fortunately, the family cat is reportedly safe but was left with a limp following the alleged incidents.

Fair punishment?

🔥 — Yes
👎 — No

😢 GPT-5.5 Has a Weird Rule About Racoons and Goblins

In the system prompt for GPT-5.5 in Codex, developers added a funny line:

“Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user's query.”

Makes you wonder how often the model brings them up...

Has your GPT ever started talking about goblins, gremlins, or raccoons?

🤣 — Yeah, it happens!
🦄 — No, but I wouldn't mind...

📱 Apple's Foldable iPad May Never Be Released

Apple's plan to create a large foldable iPad with a screen of about 20 inches might never come to market. According to a recent Bloomberg report by Mark Gurman, several sources close to the project say the device could end up as an "experimental idea" rather than a real product.

This is notable because the foldable iPad was once an important project for John Ternus, who is expected to become Apple's CEO on September 1. However, the company may decide not to release the device at all.

👨‍💻 Almost All Game Studios Use AI, But Are Afraid to Admit It

Almost all modern games already use neural networks in their development, although gamers probably don't notice it, says Jack Buser, head of gaming at Google Cloud. According to him, 9 out of 10 studios surveyed by the company admitted to using AI in their work.

However, large-scale industry research shows a much lower figure: between 40% and 50%. Buser explains that developers simply aren't ready to admit they use AI openly.

💰 Studios are practically forced to use AI to save costs. According to Buser, the cost of game development across the industry has almost doubled since 2017. AI models, while still a controversial technology, help with "heavy, repetitive, and low-value" work.

Would you play a game made with AI?

❤️ — Yes, no problem
🔥 — No, that's slop

🤩 Girl Finds Rare Endangered Axolotl In Welsh River

A 10-year-old girl discovered a rare Mexican axolotl while exploring a river in Bridgend, Wales.

According to experts, this is the first documented discovery of an axolotl in the wild in the UK, a species with only 50 to 1,000 left globally. It is believed that a former pet owner illegally released the animal.

After rescuing the injured creature, the family received expert approval to keep the axolotl, which they have named Dippy.

🐳 DeepSeek-V4 Is Here

DeepSeek-V4 is out now, and it comes in two versions: Pro and Flash. Both have a 1 million token context.

✅ Pro is the best open model for coding and agent tasks. It knows a lot about the world, and its answer accuracy is only behind Gemini 3.1 Pro.

✅ Flash is almost as good as the flagship, but it's smaller, faster, cheaper, and more efficient.

➡️ You can try it for free right now in DeepSeek Chat.

Do you use DeepSeek?

❤️ — Yes
🔥 — No