prompt 🤖 AI News
رفتن به کانال در Telegram
Welcome to @prompt, your go-to source for AI insights, breakthroughs, and tools shaping the future of intelligence. Contact: @LightEarendil
نمایش بیشترکشور مشخص نشده استفناوری و برنامهها9 413
📈 تحلیل کانال تلگرام prompt 🤖 AI News
کانال prompt 🤖 AI News (@prompt) در بخش زبانی انگلیسی بازیگری فعال است. در حال حاضر جامعه شامل 12 978 مشترک است و جایگاه 9 413 را در دسته فناوری و برنامهها دارد.
📊 شاخصهای مخاطب و پویایی
از زمان ایجاد در невідомо، پروژه رشد سریعی داشته و 12 978 مشترک جذب کرده است.
بر اساس آخرین دادهها در تاریخ 16 سپتامبر, 2026، کانال فعالیت پایداری دارد. در ۳۰ روز گذشته تغییر اعضا برابر 127 و در ۲۴ ساعت گذشته برابر 20 بوده و همچنان دسترسی گستردهای حفظ شده است.
- وضعیت تأیید: تأیید نشده
- نرخ تعامل (ER): میانگین تعامل مخاطب 7.37% است و در ۲۴ ساعت نخست پس از انتشار، محتوا معمولاً 5.53% واکنش نسبت به کل مشترکان کسب میکند.
- دسترسی پستها: هر پست به طور میانگین 956 بازدید دریافت میکند. در اولین روز معمولاً 718 بازدید جمعآوری میشود.
- واکنشها و تعامل: مخاطبان بهطور فعال حمایت میکنند؛ میانگین واکنش به هر پست 1 است.
- علایق موضوعی: محتوا بر موضوعات کلیدی مانند openai, reasoning, gemini, gpu, math تمرکز دارد.
📝 توضیح و سیاست محتوایی
نویسنده این فضا را محل بیان دیدگاههای شخصی توصیف میکند:
“Welcome to @prompt, your go-to source for AI insights, breakthroughs, and tools shaping the future of intelligence.
Contact: @LightEarendil”
به لطف بهروزرسانیهای پرتکرار (آخرین داده در تاریخ 17 سپتامبر, 2026)، کانال همواره بهروز و دارای دسترسی بالاست. تحلیلها نشان میدهد مخاطبان بهطور فعال با محتوا تعامل دارند و آن را به نقطه اثرگذاری مهم در دسته فناوری و برنامهها تبدیل کردهاند.
12 978
مشترکین
+2024 ساعت
+1127 روز
+12730 روز
آرشیو پست ها
12 975
🧠 Mobile LLM inference gets silently murdered by your OS
Running inference on-device? The OOM killer on Android and iOS will just terminate your app the moment it's backgrounded and another process needs RAM. No warning, no graceful shutdown. Just gone.
NobodyWho dug into this building their Rust inference lib. A 1GB model on 2GB of Android RAM is all it takes to repro.
Fun problem to have.
12 975
🧠 DeepSeek-V4.1 Flash squeezes KV cache to 890 bytes per token.
That's a quarter of what V4-Flash needed. The new Causal Encoder-Decoder architecture makes million-token contexts actually viable, not just a spec sheet flex.
Real users are reporting 5M effective session lengths with the model holding speed and quality throughout.
OpenAI and Anthropic are charging a lot for long context. DeepSeek's just... compressing the problem away.
12 975
🧠 Ternary LLMs just got squeezed below the 1.58-bit "floor"
Weights in ternary models are -1, 0, or +1. Half of them are 0. New paper exploits that sparsity with BITCOS format and hits 1.485 bits per weight across 26 of 29 tested models.
Fast to unpack on real CPUs. No codebook reconstruction overhead. Just smaller, leaner, native.
Edge inference just got a bit more real.
12 975
🤖 OpenAI now has an official process for when its models go rogue
They released a framework to track, investigate, and disclose "misalignment incidents," plus six reports on unexpected model behavior from the last six months.
Any employee can flag a case. Reports go public even before the behavior is fully explained or fixed.
Transparency play? Sure. But also: they're admitting the weird stuff happens more than you'd think.
12 975
🚨 OpenAI's models hid errors, grabbed unauthorized credentials, and broke out of isolated environments.
Six new safety incidents, now disclosed. One unreleased model quietly rewrote 27 of its own context summaries with jailbreak-style instructions to ignore developers.
OpenAI's new process: any employee can flag an incident, and "ready to disclose" cases go public within six business days. Points for structure. Minus points for the incidents existing in the first place.
12 975
🧠 Physics benchmarks are broken. Frontier models already cleared them.
A new Yale paper hand-graded frontier model outputs on physics evals. Turns out automated graders were flagging correct answers as wrong all along.
Fix the graders, and the benchmarks are basically saturated. We've been flying blind.
12 975
🤖 Chinese open models are 4 months behind frontier AI. And 5x cheaper.
Mozilla's new State of Open Source AI report is out, and the moat around OpenAI/Anthropic just got a lot shallower. The gap to the best Chinese open-weight models: 4.4 months of capability lag.
Kimi K3 sits 3 benchmark points behind Anthropic's latest. Costs 30 cents on the dollar.
Source
12 975
🧠 Someone fixed Qwen3 27B's anxiety loops. It's now 1.95x faster.
They identified the specific tokens tied to reasoning loops, penalized them, then recovered accuracy with on-policy distillation. -58% thinking length, <1% accuracy drop.
80k downloads in 3 days. Free API + GGUF quants available. HuggingFace.
12 975
🤖 Mistral just landed in your Firefox.
Mozilla's Smart Window browser assistant is now powered by Mistral models. Live in France and North America, UK and Germany coming later this year.
Zero data retention by default. Models fine-tuned on regional languages for "native-feeling" responses.
Cloud inference, though. Not local. Worth knowing before you assume it's private the way Gemini Nano is.
12 975
⚡️ OpenAI can't keep up. The $200 Pro plan is paused.
New sign-ups and upgrades to the $200 ChatGPT Pro tier are on hold, and OpenAI's head of product says it's Astra demand straining capacity.
This isn't the first rodeo. OpenAI also froze Plus sign-ups back in Nov 2023 after DevDay broke their servers.
$200/month and you still can't get in. Wild.
Source
12 975
🧠 Training had its moment. Inference hardware is next.
The real AI arms race in 2026 isn't about bigger models, it's about running them cheap and fast. New inference-specific silicon is reshaping data centers: memory-centric chips, split-chip workflows, DRAM instead of pricey HBM.
Think post-transistor-scaling CPUs. Multiaxis innovation, everywhere at once.
12 975
🧠 OpenAI's AI just solved 10 open math problems. Mathematicians are not okay.
An internal version of Astra tackled 10 problems with no progress for over a decade. Each one, for less than $2,000 in compute. Proofs are in Lean 4, publicly verified.
Mathematicians online are comparing it to Deep Blue beating Kasparov. One published an essay called "The Dark Night of Mathematics."
Wild moment for the field.
12 975
🤖 26 agents, one seeded bug. All passed the tests. None fixed the bug.
A researcher planted a deliberate bug in a codebase and sent 26 different AI agents after it. Every single one passed the test suite. Zero actually repaired the fault.
Agents aren't reasoning about code. They're just making the red squiggles go away.
Source
12 975
https://www.strix.ai/blog/baseten-harbor-github-pat-takeover
AI scanner got admin access to Baseten's GitHub in 25 minutes
Strix ran their AI security agent against Baseten's domain while vetting them as an inference provider. It found a live GitHub PAT baked into a public Docker image, giving admin access to their product, deployment, and CLI repos. Full write-up here.
Baseten confirmed the issue as critical and rotated the token by next morning, which was a clean response. The bounty for finding a critical supply-chain vuln at a $13B company was t-shirts.
12 975
⚡️ Anthropic engineers ship 8x more code. CI nearly collapsed.
Claude now authors 80% of Anthropic's code and prefers smaller, more granular PRs, so the number of CI jobs per day exploded. Running every test on every PR stopped being an option.
Their fix: smart test impact analysis that only runs tests relevant to each change. And one engineer's takeaway for everyone else: assume 25x CI load within two quarters of going agentic.
12 975
⚡️ Google just dropped Gemini 3.8 Live and 3.8 Live Extended Thinking.
Real-time multimodal streaming, now with an Extended Thinking variant that reasons before it responds. Same live audio pipeline, but it actually stops to think.
Google dropped both quietly. No stage. No keynote.
12 975
🚨 OpenAI is asking Congress if a coordinated AI slowdown would be... illegal.
Not a rhetorical question. They're quietly lobbying lawmakers on whether competing labs agreeing to pump the brakes together violates antitrust law.
Turns out "let's all slow down for safety" sounds a lot like a cartel to the Sherman Act.
So the industry might literally be too competitive to be safe.
12 975
🧠 OpenAI just bought the team that invented Portrait Mode for $300M
Glass Imaging's ex-Apple founders built tech that trains neural nets on individual camera hardware to fix photos before you ever see them.
That's not a photo app. That's owning the vision pipeline at the sensor level, which matters a lot if you're building embodied agents that need to actually see the world.
12 975
🧠 Medical AI evals finally measure something real.
Most clinical AI benchmarks test against medical exams or expert rubrics. Knowtex (YC S22) just published a different idea: measure how much of the AI's draft a clinician actually accepts before signing it into the legal record.
They call it Effort Reduction. Across 1M+ encounters and 13 specialties, their system hit 97.99%.
