UtroAI
BREAKING AI FRONTIER
YouTube
Wait... Smart Glasses are Suddenly Good?
Trying the Meta Rayban Display glasses
Trying the Orion prototypes: https://youtu.be/G0eKzU_fV00
MKBHD Merch: http://shop.MKBHD.com
Playlist of MKBHD Intro music: https://goo.gl/B3AWV5
~
http://twitter.com/MKBHD
http://instagram.com/MKBHD
http://facebook.com/MKBHD
Trying the Orion prototypes: https://youtu.be/G0eKzU_fV00
MKBHD Merch: http://shop.MKBHD.com
Playlist of MKBHD Intro music: https://goo.gl/B3AWV5
~
http://twitter.com/MKBHD
http://instagram.com/MKBHD
http://facebook.com/MKBHD
UtroAI
ai-2027.com
YouTube
Robot Plumbers, Robot Armies, and Our Imminent A.I. Future | Interesting Times with Ross Douthat
Is artificial intelligence about to take your job? According to Daniel Kokotajlo, the executive director of the A.I. Futures Project, that should be the least of your worries. Kokotajlo was once a researcher for OpenAI, but left after losing confidence in…
Please open Telegram to view this post
VIEW IN TELEGRAM
UtroAI
Two Minute Papers 🔗 Alfa Evolve AGI
YouTube
Tucker Debates Bryan Johnson on His Methods of Anti-Aging and Ethics of Living Forever
@BryanJohnson is a very smart, very rich, very well-meaning man who wants to live forever. That sounds like a terrible idea. This is one of the most interesting debates we've ever had.
Watch more here: https://bit.ly/3RCq6cc
Subscribe to the new Tucker…
Watch more here: https://bit.ly/3RCq6cc
Subscribe to the new Tucker…
AI heads to orbit in November 2025 | Moment X
NVIDIA and startup Starcloud are sending a data‑center‑class GPU (NVIDIA H100) into space. On board, they plan to run Google’s open Gemma model to prove that large language models can operate off‑planet.
“In space, you get almost unlimited, low‑cost renewable energy.” — Philip Johnston, Starcloud cofounder & CEO.
Why it matters
• Power & cooling: direct solar + the vacuum of space as a natural heat sink → no water‑intensive cooling.
• Impact: Starcloud projects ~10× lower energy costs and lifecycle CO₂ savings vs. Earth‑based data centers (even accounting for the launch).
• Use cases: on‑satellite processing of Earth‑observation streams (optical, hyperspectral, and SAR—~10 GB/s) could cut wildfire or distress‑signal detection from hours to minutes.
Big picture: If this works, tomorrow’s data centers may live in orbit. Security & governance questions remain open.
Source: NVIDIA blog.
Comment: People are racing to build energy‑independent, self‑sustaining, spatially distributed AI. We’ll be watching.
NVIDIA and startup Starcloud are sending a data‑center‑class GPU (NVIDIA H100) into space. On board, they plan to run Google’s open Gemma model to prove that large language models can operate off‑planet.
“In space, you get almost unlimited, low‑cost renewable energy.” — Philip Johnston, Starcloud cofounder & CEO.
Why it matters
• Power & cooling: direct solar + the vacuum of space as a natural heat sink → no water‑intensive cooling.
• Impact: Starcloud projects ~10× lower energy costs and lifecycle CO₂ savings vs. Earth‑based data centers (even accounting for the launch).
• Use cases: on‑satellite processing of Earth‑observation streams (optical, hyperspectral, and SAR—~10 GB/s) could cut wildfire or distress‑signal detection from hours to minutes.
Big picture: If this works, tomorrow’s data centers may live in orbit. Security & governance questions remain open.
Source: NVIDIA blog.
Comment: People are racing to build energy‑independent, self‑sustaining, spatially distributed AI. We’ll be watching.
Perplexity Opens
the “Patent Gates”
Patent search — until now the domain of expensive attorneys and specialists fluent in Boolean incantations like (keyword1 OR keyword2) AND (inventor:musk). A whole science of its own.
Word of the day: patentoscope.
The magic: you can now ask in natural language — e.g. “What patents were filed in 2025 in [topic]?” — and instantly get
a) summaries based on real patent reviews, and
b) direct links to the patent docs.
Superpower of the
Week · Perplexity
1 Patents: from shamanism to plain language
Now you ask like a human — get curated, relevant patents with key takeaways, viewer, and links to originals.
It works as a dialogue: refine on the fly, the system grasps synonyms and adjacent terms, and, when needed, fetches articles or public code as prior art.
Free in beta; Pro / Max unlock higher quotas.
How to Use Today (3 steps)
– Ask: “Key patents of 2025 for [topic]?” / “Any patents covering [this idea]?”
– Compare: “Compare these two patents for novelty and claims.”
– Map: “Build a table: number → applicant → core idea.”
Vector (Where It’s Headed)
Δ Barrier ↓ · Δ Latency ↓ · Δ Coverage ↑ · Δ Programmability ↑
The unit of IP knowledge has shifted from operator syntax to engineer’s question.
That collapses the cycle “question → map → solution” from weeks to hours — and injects design-around thinking at the earliest stage.
the “Patent Gates”
Patent search — until now the domain of expensive attorneys and specialists fluent in Boolean incantations like (keyword1 OR keyword2) AND (inventor:musk). A whole science of its own.
Word of the day: patentoscope.
The magic: you can now ask in natural language — e.g. “What patents were filed in 2025 in [topic]?” — and instantly get
a) summaries based on real patent reviews, and
b) direct links to the patent docs.
Superpower of the
Week · Perplexity
1 Patents: from shamanism to plain language
Now you ask like a human — get curated, relevant patents with key takeaways, viewer, and links to originals.
It works as a dialogue: refine on the fly, the system grasps synonyms and adjacent terms, and, when needed, fetches articles or public code as prior art.
Free in beta; Pro / Max unlock higher quotas.
How to Use Today (3 steps)
– Ask: “Key patents of 2025 for [topic]?” / “Any patents covering [this idea]?”
– Compare: “Compare these two patents for novelty and claims.”
– Map: “Build a table: number → applicant → core idea.”
Vector (Where It’s Headed)
Δ Barrier ↓ · Δ Latency ↓ · Δ Coverage ↑ · Δ Programmability ↑
The unit of IP knowledge has shifted from operator syntax to engineer’s question.
That collapses the cycle “question → map → solution” from weeks to hours — and injects design-around thinking at the earliest stage.
Sources: Perplexity Patents (official release + blog), Axios (first look), The Verge (news), Perplexity Research (technical note), @serge_ai.