Generative AI Is Altering All the things. However What’s Left When the Hype Is Gone?
Will Douglas Heaven | MIT Know-how Assessment
“The thrilling reality is, we don’t actually know. For whereas artistic industries—from leisure media to vogue, structure, advertising, and extra—will really feel the influence first, this tech will give artistic superpowers to everyone. In the long term, it may very well be used to generate designs for nearly something, from new kinds of medicine to garments and buildings. The generative revolution has begun.”
New ‘Mobile Glue’ Idea Might Heal Wounds, Regrow Nerves
Monisha Ravisetti | CNET
“Researchers from the College of California, San Francisco introduced a captivating innovation on Monday. They name it ‘mobile glue’ and say it may at some point open doorways to huge medical achievements, like constructing organs in a lab for transplantation and reconstructing nerves that’ve been broken past the attain of normal surgical restore.”
The Viral AI Avatar App Lensa Undressed Me—With out My Consent
Melissa Heikkilä | MIT Know-how Assessment
“Lensa generates its avatars utilizing Secure Diffusion, an open-source AI mannequin that generates photos based mostly on textual content prompts. Secure Diffusion is constructed utilizing LAION-5B, a large open-source information set that has been compiled by scraping photos off the web. And since the web is overflowing with photos of bare or barely dressed ladies, and footage reflecting sexist, racist stereotypes, the info set can be skewed towards these sorts of photos.”
Scientists Could Have Discovered the First Water Worlds
John Timmer | Ars Technica
“…continued remark has produced information that signifies the planets are a lot much less dense than we initially thought. And the one practical solution to get the form of densities they now appear to have is for a considerable quantity of their quantity to be occupied by water or an analogous fluid. We do have our bodies like this in our Photo voltaic System—most notably the moon Europa, which has a rocky core surrounded by a watery shell capped by ice. However these new planets are a lot nearer to their host star, which suggests their surfaces are in all probability a blurry boundary between an unlimited ocean and a steam-filled environment.”
The Tech Is Lastly Good Sufficient for an Airship Revival
Michael Koziol | IEEE Spectrum
“At Moffett Subject in Mountain View, Calif., Lighter Than Air (LTA) Analysis is floating a brand new method to a expertise that noticed its rise and fall a century in the past: airships. Though airships have lengthy since been supplanted by planes, LTA, which was based in 2015 by CEO Alan Weston, believes that by way of a mixture of latest supplies, higher building methods, and technological developments, airships are poised to—not reclaim the skies, actually—however discover a new area of interest.”
The Actual Fusion Power Breakthrough Is Nonetheless Many years Away
Gregory Barber | Wired
“Right now, the NIF researchers stated they bought as a lot vitality out as their laser fired on the experiment—a large, long-awaited achievement. However the issue is that the vitality in these lasers represents a tiny fraction of the whole energy concerned in firing up the lasers. By that measure, NIF is getting manner lower than it’s placing in. ‘That kind of breakeven is manner, manner, manner, manner down the street,’ Cappelli says. ‘That’s a long time down the street. Perhaps even a half-century down the street.’i”
AI-Generated Faux Faces Have Grow to be a Hallmark of On-line Affect Operations
Shannon Bond | NPR
“Fb guardian firm Meta says greater than two-thirds of the affect operations it discovered and took down this 12 months used profile footage that have been generated by a pc. As the unreal intelligence behind these fakes has grow to be extra extensively out there and higher at creating life-like faces, dangerous actors are adapting them for his or her makes an attempt to govern social media networks.”
What Does It Imply to Align AI With Human Values?
Melanie Mitchell | Quanta
“We people are liable to giving machines ambiguous or mistaken directions, and we would like them to do what we imply, not essentially what we are saying. …To resolve this downside, [AI researchers] imagine, we should discover methods to align AI techniques with human preferences, objectives and values. …However with no higher understanding of what intelligence is and the way separable it’s from different points of our lives, we can’t even outline the issue, a lot much less discover a answer. Correctly defining and fixing the alignment downside gained’t be straightforward; it is going to require us to develop a broad, scientifically based mostly idea of intelligence.”
Picture Credit score: Clark Van Der Beken / Unsplash