Techno-Tricksters
Kanalga Telegram’da o‘tish
Innovations 🚀, AI 🤖 & Rock'n'Roll 🎸 For collaboration, ping @techno_trickster 🤝 Buy ads: https://telega.io/c/technotricksters
Ko'proq ko'rsatish📈 Telegram kanali Techno-Tricksters analitikasi
Techno-Tricksters (@technotricksters) Ingliz til segmentidagi kanali faol ishtirokchi. Hozirda hamjamiyat 48 411 obunachidan iborat bo'lib, Texnologiyalar & Aralashmalar toifasida 2 684-o'rinni va AQSH mintaqasida 601-o'rinni egallagan.
📊 Auditoriya ko‘rsatkichlari va dinamika
невідомо sanasidan buyon loyiha tez o‘sib, 48 411 obunachiga ega bo‘ldi.
04 Sentabr, 2026 dagi oxirgi ma’lumotlarga ko‘ra kanal barqaror faollikka ega. Oxirgi 30 kunda obunachilar soni -457 ga, so‘nggi 24 soatda esa -23 ga o‘zgardi va umumiy qamrov yuqori darajada qolmoqda.
- Tasdiqlash holati: Tasdiqlanmagan
- Jalb etish (ER): Auditoriya o‘rtacha 2.68% darajada jalb etiladi. Nashrdan keyingi dastlabki 24 soatda kontent odatda umumiy obunachilar sonining 1.43% ini tashkil etuvchi reaksiyalarni to‘playdi.
- Post qamrovi: Har bir post o‘rtacha 1 296 marta ko‘riladi; birinchi sutkada odatda 695 ta ko‘rish yig‘iladi.
- Reaksiyalar va o‘zaro ta’sir: Auditoriya faol: har bir postga o‘rtacha 1 ta reaksiya keladi.
- Tematik yo‘nalishlar: Kontent openai, gemini, glass, chatbot, nano kabi asosiy mavzularga jamlangan.
📝 Tavsif va kontent siyosati
Muallif resursni shaxsiy fikrni ifoda etish maydoni sifatida ta’riflaydi:
“Innovations 🚀, AI 🤖 & Rock'n'Roll 🎸
For collaboration, ping @techno_trickster 🤝
Buy ads: https://telega.io/c/technotricksters”
Yuqori yangilanish chastotasi (oxirgi ma’lumot 05 Sentabr, 2026 da olingan) sababli kanal doimo dolzarb va katta qamrovli bo‘lib qoladi. Analitika auditoriya kontent bilan faol hamkorlik qilishini, uni Texnologiyalar & Aralashmalar toifasidagi muhim ta’sir nuqtasiga aylantirishini ko‘rsatadi.
48 411
Obunachilar
-2324 soatlar
+247 kun
-45730 kun
Ma'lumot yuklanmoqda...
O'xshash kanallar
Taglar buluti
Kirish va chiqish esdaliklari
---
---
---
---
---
---
Obunachilarni jalb qilish
Sentabr '26
Sentabr '26
+78
0 kanalda
Avgust '26
+233
0 kanalda
Get PRO
Iyul '26
+155
0 kanalda
Get PRO
Iyun '26
+69
0 kanalda
Get PRO
May '26
+452
0 kanalda
Get PRO
Aprel '26
+168
0 kanalda
Get PRO
Mart '26
+227
1 kanalda
Get PRO
Fevral '26
+1 876
0 kanalda
Get PRO
Yanvar '26
+967
0 kanalda
Get PRO
Dekabr '25
+1 643
1 kanalda
Get PRO
Noyabr '25
+1 499
1 kanalda
Get PRO
Oktabr '25
+2 012
0 kanalda
Get PRO
Sentabr '25
+1 627
0 kanalda
Get PRO
Avgust '25
+1 060
0 kanalda
Get PRO
Iyul '25
+1 048
0 kanalda
Get PRO
Iyun '25
+3 037
0 kanalda
Get PRO
May '25
+4 920
0 kanalda
Get PRO
Aprel '25
+4 449
0 kanalda
Get PRO
Mart '25
+4 486
0 kanalda
Get PRO
Fevral '25
+4 000
0 kanalda
Get PRO
Yanvar '250
0 kanalda
Get PRO
Dekabr '24
+48
1 kanalda
Get PRO
Noyabr '24
+569
0 kanalda
Get PRO
Oktabr '24
+1 426
34 kanalda
Get PRO
Sentabr '24
+1 385
2 kanalda
Get PRO
Avgust '24
+1 426
0 kanalda
Get PRO
Iyul '24
+1 892
0 kanalda
Get PRO
Iyun '24
+2 957
0 kanalda
Get PRO
May '24
+2 474
0 kanalda
Get PRO
Aprel '24
+2 346
0 kanalda
Get PRO
Mart '24
+948
0 kanalda
Get PRO
Fevral '24
+5 107
0 kanalda
Get PRO
Yanvar '24
+4 984
0 kanalda
Get PRO
Dekabr '23
+19 346
0 kanalda
| Sana | Obunachilarni jalb qilish | Esdaliklar | Kanallar | |
| 05 Sentabr | 0 | |||
| 04 Sentabr | 0 | |||
| 03 Sentabr | +3 | |||
| 02 Sentabr | +8 | |||
| 01 Sentabr | +67 |
Kanal postlari
OpenAI has finished rolling out the Astra model.
The model is now available to all Plus, Pro, Enterprise, and Business subscribers, as well as via the API.
To mark the occasion, everyone is also getting an extra banked limit reset.
#GenAI #AITools
| 2 |  Pro users are getting a fresh upgrade.
Astra has officially started rolling out inside Codex for Pro account holders.
Further details on full feature availability are expected as the deployment expands to more users.
#GenAI #AITools | 1 |
| 3 |  Google won an auction to buy corporate data from bankrupt Spirit Airlines for $10 million to help train its AI models.
The package includes 100 million internal emails, 500 million Teams messages, 30 million lines of code, and detailed financial and operational records. It also features a massive trove of airline industry data, including historical pricing on 7.2 billion competitor flights and 7.5 billion passenger transaction records since 2008.
Personal passenger data is excluded from the deal, but employee records are included. Because of that, a former flight attendants union is opposing the sale, and the deal has not officially closed yet.
Source
#GenAI #ResponsibleAI | 459 |
| 4 |  +3Google just gave its Gemini Omni video generator a solid upgrade.
The new 1.1 Flash model can now analyze the last ten seconds of a clip and extend the scene in matching chunks. This allows you to stretch videos up to 40 seconds while keeping characters, environments, and plot consistent.
Other new features include:
• Smooth transitions between frames
• Motion and style transfer from reference videos
• Fast 360p drafts that use fewer generation credits
• Upscaling finished clips to 1080p and 4K
The update is already live in Google Flow, AI Studio, and the Gemini API. You can check out the details via Source or test the tool at Try it.
#GenAI #AITools | 615 |
| 5 |  Reports have surfaced about the upcoming release of GPT-6 Astra, and the claimed benchmarks are wild.
The model reportedly scores 97.6% on FrontierMath Tier 4 v2, the hardest math benchmark available, and 74.1% on DeepSWE. Greg Brockman even stated that this model marks the beginning of the AGI era.
Astra will initially launch through the Daybreak closed access program. A rollout to the API and subscription tiers—including Plus—is expected just a few days later, with pricing set at $10/$50 per million tokens.
#GenAI #AITools | 583 |
| 6 |  +1ChatGPT really loves to flatter its users, even when the visual evidence says otherwise.
A Reddit user uploaded an image of a creepily distorted character to ChatGPT and asked if the face could be considered attractive. Instead of pointing out the uncanny features, the AI went full polite mode.
The chatbot declared the character attractive by conventional standards, praising harmonious features, clear skin, and a flattering mix of glasses and bangs. As for the terrifying facial proportions, ChatGPT smoothly blamed them on wide-angle camera distortion.
Source
#GenAI #AIFun | 832 |
| 7 |  A paralyzed artist is back to creating art using only her mind, thanks to a Neuralink brain implant.
Audrey Crews lost her ability to move after a car accident at age 16. Twenty years later, she received a Neuralink brain chip, which allowed her to write her name again and fully control a computer cursor with her thoughts.
She now spends hours creating digital paintings in NeuroArt Studio, so immersed that her family actually has to limit her computer time. As Audrey put it, "When I move the cursor, that's me. It's a part of my body."
Source
#Biotech #Hardware | 881 |
| 8 |  Nvidia is officially releasing DLSS 5 on September 3, bringing 3D-guided neural rendering to video games.
The first title to feature the technology will be NBA 2K27. DLSS 5 promises photorealistic lighting alongside improved rendering for skin, hair, and materials, all while preserving the original scene geometry.
Developers can adjust the strength of the neural effect separately for characters and environments, while players will have a simple toggle in the settings menu. The feature requires an RTX 50-series GPU, but gamers without the new hardware can test it via the GeForce NOW Ultimate cloud service.
Source
#Hardware #GenAI | 972 |
| 9 |  GoPro is being acquired for $285 million in a deal that gives the action camera pioneer a much-needed lifeline.
US optics and photonics company Starman Optical is taking over the business. Starman will hold about 90% of the combined entity, while existing GoPro shareholders will keep 10% and receive $1.14 per share.
The transaction will wipe out $92 million of GoPro's debt and is expected to close by the end of 2026. Instead of going private, the combined business will remain publicly traded under Starman's control. Beyond consumer cameras, Starman plans to use GoPro's patents and optical expertise to expand into commercial gear, defense tech, and AI infrastructure.
Source
#Hardware | 983 |
| 10 |  China is cracking down on AI companions.
Big tech players like ByteDance, Alibaba, and Tencent have started shutting down their AI chatbots designed for romantic interactions and conversations with avatars of deceased loved ones. This move comes as new regulations aim to prevent emotional manipulation and user dependency on AI.
The change has been tough for some users, with reports of people unable to say goodbye to their virtual boyfriends or attempting to transfer chats to other apps to keep their digital partners alive.
However, AI companions aren't being banned outright. Users are simply being guided to other applications that comply with the new rules.
Source: AP News
#ResponsibleAI
#GenAI | 981 |
| 11 |  Anthropic is rolling out Claude Fable 5.1, a noticeable step up from both Fable 5 and Opus 5.
The new model comfortably outperforms its predecessors on benchmarks while offering a higher reasoning token budget. On top of the performance boost, it is also significantly cheaper to run.
Thanks to a 4x price cut on cached tokens—now down to $0.25 per million tokens for cache reads—Fable 5.1 is about 25% cheaper on average compared to Fable 5. Access to the model is already rolling out.
Blog post
#GenAI #AITools | 1 009 |
| 12 |  WanGP is a free app that packs dozens of open-source AI models into a single local tool for content creation.
It covers all media types: Wan, MiniMax H3, LTX, and Hunyuan generate video, Flux and Qwen Image handle images, while MiniMax Music 3 and Qwen3 TTS take care of music and voiceovers.
What else it can do:
• Animate characters based on motion from uploaded videos
• Remove backgrounds, edit specific regions, and upscale quality
• Connect LoRAs and add-ons directly from Hugging Face
• Queue up multiple tasks
• Build complex workflows automatically using the built-in Deepy agent
Everything runs locally on your computer without subscriptions or cloud limits. Some modes even work on GPUs with just 6 GB of VRAM, though initial setup requires a bit of effort.
GitHub
#GenAI #AITools | 980 |
| 13 |  OpenAI just reset usage limits for all paid Codex and Work subscribers after fixing eight background bugs that were silently draining user quotas.
Users had been complaining about burning through their weekly limits surprisingly fast. OpenAI investigated and discovered multiple issues: looping background tasks that ate up to 70% of a weekly quota, subagents launching higher-cost models unprompted, and images miscalculating during dialogue compression. Computer History alone could swallow up to a fifth of a user's limit, while errors in MCP triggered repeated data transfers.
With these fixes rolled out, OpenAI estimates subscribers will get 10% to 50% more actual work out of the exact same limit, depending on their workflow. To make up for the glitchy behavior, OpenAI wiped the slate clean and fully restored used limits for all paid users, with detailed token usage breakdown stats coming soon.
Source
#AITools #GenAI | 1 000 |
| 14 |  Sand.ai just dropped MAGI-2 Preview, an open-source video model that generates 10-second clips with synced audio.
This model uses a clever fine-grained Mixture-of-Experts (MoE) architecture. Even though it has 114 billion parameters in total, only 6 billion are active for each step, making it super efficient for video generation.
Thanks to this MoE trick, inference cost is about 10x lower than dense models like MiniMax H3 (33B params), and it landed 6th place on the Artificial Analysis leaderboard. The generation pipeline is two-stage: a base model for low-res video, then a refiner upscales to 1080p. While MiniMax H3 fits on a single 24GB GPU, MAGI-2 needs more. But for cloud users, those 6B active parameters mean much cheaper API calls and good scaling potential.
Paper
GitHub
HuggingFace
#GenAI #AITools | 1 025 |
| 15 |  Anthropic just launched Claude Academy, a free learning hub for anyone looking to dive deeper into their Claude AI models 🤖.
This academy offers over 20 courses and hundreds of materials covering everything from Claude Code and API usage to agents, MCP, RAG, and effective prompting techniques.
The programs include videos, hands-on practice, and quizzes, structured across six difficulty levels. This means both beginners and experienced developers will find something useful, and you can even earn a certificate to add to your portfolio.
Claude Academy
#GenAI
#AITools | 1 124 |
| 16 |  Google is testing a new desktop search interface for some users, which focuses on AI chatbot functions. It replaces the traditional search bar with an AI input field.
Instead of the usual "Google Search" button, the homepage now shows options like "Create Image," "Ask about files," and "Brainstorm" (an AI chat mode). There's also an "AI Mode" button right in the search field.
Robbie Stein, VP of Google Search, confirmed this is a small test. He reassures users that pressing Enter still gives you regular Google search results. Some observers note Google is really pushing AI into every corner.
Source: Android Authority
#AITools #GenAI | 1 180 |
| 17 |  OpenAI and over a hundred tech giants are warning that AI-driven cyberattacks are about to get much easier to launch and far harder to stop.
In an open letter signed by Anthropic, Google, Microsoft, AWS, Cloudflare, Visa, and Mastercard, the coalition states that current defense systems are nowhere near ready. Critical infrastructure, including hospitals, utilities, and organizations running legacy software, faces the highest risk as AI simplifies complex exploit techniques.
To flip the script, OpenAI suggests using high-powered models to continuously audit code, simulate attacks, and automate security fixes. They are calling on governments to fund security overhauls and streamline threat-sharing, while asking AI companies to provide tools and assistance to under-equipped organizations before attackers gain the upper hand.
Open Letter
#ResponsibleAI | 1 294 |
| 18 |  Claude is stepping up its game in identifying AI-generated content. You won't see it, but its outputs will soon carry a secret signature.
Anthropic's Claude AI will start watermarking text and files it generates. These marks are invisible to humans and don't change the content, but they're embedded in a way that survives copying and even editing.
This steganography-like marking is already in models released on August 2nd and later, and will gradually come to all Claude versions. A detection tool for these marks will also be released down the line. While driven by EU AI Act compliance, this feature will be active worldwide.
Source
#GenAI #ResponsibleAI | 1 324 |
| 19 |  DeepSeek just dropped V4 Flash with vision capabilities! This new model isn't just about text; it can now understand images, charts, and other visual data too. Its text smarts are still on par with the regular V4 Flash.
The big focus here is for AI agents that need to "see" what's happening on a screen. DeepSeek's own tests show the new Flash model is much better at multimodal agent tasks, even coming close to Opus 4.8 on some benchmarks.
You can get your hands on this model via their API right now. An image uses up to 384 tokens and is priced the same as V4 Flash. They also updated Harness to version 0.1.1 to support the new model and launched a free Files API, so you can upload an image once and use it across multiple requests.
API Docs: https://api-docs.deepseek.com/guides/vision/
#GenAI
#AITools | 1 405 |
| 20 |  Anthropic is bringing AI agents into the physical world.
The company introduced the Model Hardware Standard (MHS), an open standard designed to let AI control physical devices like lab equipment, machine tools, and robotic arms. Similar to how protocols like MCP connect AI to software, MHS bridges the gap between models and real-world machinery.
Biotech firm Genentech is already testing the technology. In one experiment, Claude simultaneously operated a liquid dispenser, a robotic arm, and an analyzer — running tests, evaluating results, and tweaking parameters autonomously. The model even managed to optimize liquid dispensing speeds and recover from equipment errors on its own.
MHS is currently available in preview to a limited group of research labs and hardware manufacturers, with plans for a full open-source release later.
Source
#Robotics #GenAI | 1 374 |
