• Link & Load
  • Posts
  • 🤖 Amazon’s touch-sensitive robots | 🛡️ Google unmasks fresh Russian malware | 🧠 Mistral’s bargain-blitz AI

🤖 Amazon’s touch-sensitive robots | 🛡️ Google unmasks fresh Russian malware | 🧠 Mistral’s bargain-blitz AI

Your quick hit of feel-y bots, hacker swats, and cut-price brainpower—before the Wi-Fi gremlins wake up.

🔥 Warm-Up

Welcome back, bandwidth bandit — time to juice-up that RSS-deprived cortex.
• Amazon just gave its warehouse bots a literal sense of touch.
• Google spotted fresh Russian malware, because of course it did.
• French startup Mistral says its new AI model punches above its price class.
• A British company built an AI avatar that signs for the Deaf community.
• UK experts think therapy chatbots might need therapy themselves.

🚀 Today in Tech

Amazon’s “Vulcan” robot feels its way around your packages

Source: https://www.theverge.com/news/662452/amazon-vulcan-warehouse-robot-sense-touch
TL;DR: Amazon’s new two-armed bot uses force sensors and ML to pluck items from fabric bins, handling roughly 75 % of the company’s inventory in pilot sites. It has already processed half-a-million orders while sparing humans the worst shelf-contortions. Amazon insists Vulcan complements, not replaces, staff— though labor groups aren’t buying that kumbaya pitch. Rollouts across the US and Europe are next.

Google flags “LOSTKEYS” malware from Russia-linked Cold River crew

Source: https://www.reuters.com/technology/cybersecurity/google-identifies-new-malware-linked-russia-based-hacking-group-2025-05-07/
TL;DR: The spyware steals files and phones home system data, extending a campaign that already targeted diplomats and NGOs. Google’s Threat Analysis Group says the code is new but the tactics feel recycled: phishing, bad attachments, rinse, repeat. Advice: patch, enable 2FA, and stop opening random .ZIPs from “totally-real-embassy.ru.” Kremlin, as usual, denies everything.

Mistral’s Medium 3 model shoots for “champagne taste on seltzer money”

Source: https://techcrunch.com/2025/05/07/mistral-claims-its-newest-ai-model-delivers-leading-performance-for-the-price/
TL;DR: At $0.40 per million input tokens, the Parisian upstart claims 90 % of Anthropic Sonnet’s accuracy for a fraction of the cash. Medium 3 runs on four GPUs, slots into AWS SageMaker today, and is heading for Azure and Vertex AI next. Benchmarks show it nudging past Meta’s Llama 4 Maverick. One analyst quips that it’s “the Dacia Sandero of LLMs: cheap, surprisingly peppy, and no cup holders.”

AI sign-language avatars promise an interpreter in your pocket

Source: https://www.wired.com/story/silence-speaks-deaf-ai-signing/
TL;DR: UK startup Silence Speaks can translate text to British and American Sign Language, rendering fluid gestures with a photoreal avatar. Transport for London will trial it in stations so Deaf commuters can scan QR codes for real-time updates. On-device, real-time translation is at least a year away, but early demos already beat clunky video loops. Accessibility experts call it a big win; mime artists call their agents.

Therapy chatbots called “unsafe” by British clinicians

Source: https://www.theguardian.com/technology/2025/may/07/experts-warn-therapy-ai-chatbots-are-not-safe-to-use
TL;DR: A UK panel warns mental-health bots often miss red-flag language and can dish out harmful advice — like suggesting yoga for severe depression (downward-dog won’t cut it, mate). Regulators mull classifying them as medical devices, triggering hefty audits. One psychiatrist joked that the bots “have no couch-side manner.” On the upside, at least they don’t ask how that makes them feel. (Dad joke achieved.)

What do you think about Link & Load?

Login or Subscribe to participate in polls.

🏁 Sign-off

May your battery stay green and your inbox stay lean — see you tomorrow ⚡️

Reply

or to participate.