341 stories
·
21 followers

Saturday Morning Breakfast Cereal - AI

1 Comment and 3 Shares


Click here to go see the bonus panel!

Hovertext:
I've been told we're now down to only 4 gibberish panels.


Today's News:
Read the whole story
duerig
13 days ago
reply
This is pretty accurate. I've taken a look at 'books written by a computer' starting in the 80s and coming down to today. And it basically comes down to somebody rolling the dice until something interesting results and publishing their curated thing.
Share this story
Delete

Poison Control Centers Are Fielding A Surge Of Ivermectin Overdose Calls

2 Comments
Health experts and medical groups are pushing to stamp out the growing use of ivermectin, an anti-parasite drug, to treat COVID-19, amid warnings that it can cause harmful side effects and that there

The nation's poison control centers saw a 245% jump in reported exposure cases from July to August as more people take the anti-parasite drug that some falsely claim treats COVID-19.

(Image credit: Denis Farrell/AP)

Read the whole story
duerig
48 days ago
reply
It is so hard to wrap my head around this madness. I understand that people can be skeptical of something like the vaccines because they are new and that some combination of fraudsters and foreign agents can whip that skepticism into denialism. But how does that become transmuted into absolute certainty that they should be taking untested drugs meant only for animals?

If somebody convinced me that it was unsafe to drink milk, I would not then be more inclined to eat cat food. What is the missing link? Is it really just 'the libtards say you shouldn't eat cat food!' and that is all the convincing some people need?
Share this story
Delete
1 public comment
dreadhead
48 days ago
reply
People need to stop horsing around with this stuff.
Vancouver Island, Canada

Taking Stock Of The Great and Cowardly Press Freakout Of August 2021 | Talking Points Memo

2 Comments and 4 Shares

With the American war in Afghanistan and the American withdrawal from Afghanistan now definitively over, I’ve been trying to put the entirety of the last four weeks into some perspective. As you can see I’ve been fairly dug in on the proposition that the great majority of the criticism we’ve seen amounts to ignorance and deflection. Pulling the plug on a failed or misconceived mission isn’t pretty. But it is inevitable. The ugliness is built into the failure rather than a consequence of recognizing it. Most of what we’ve seen is an attempt to deny the failure (mostly hawks) or imagine that withdrawing would be orderly and free of consequences. But with all this reasoning, what parts were handled poorly? What could have been better organized or cleaner?

Perhaps this is only a matter of stepping back and with the benefit of clear eyes and more perspective and, well … agreeing with myself. But honestly, I think I really agree with myself. The airlift evacuation appears to have transported well over 110,000 people out of the country, an astonishing feat under any circumstances and probably unprecedented for a civilian airlift in a kinetic military context and in the context of state collapse.

What happened two weeks ago was that the US-backed government fell. Quickly. And the US President, who had decided to end the US mission in Afghanistan without conditions, allowed it to fall rather than changing his mind. That is the entirety of what happened. Shifting gears to prevent the collapse would have signaled not only that the decision was wrong but also was poorly arrived at in the first place since the collapse of the government was always the probable and most likely the inevitable result of the decision.

Most Monday-morning quarterbacking of the “failure of execution” school doesn’t posit that the government would have survived, only that it might have lasted months or maybe a year longer, thus allowing the US to pass off the messiness on someone else. In other words, the failure of execution was largely a matter of optics. Extend the country’s civil war for a few more months or years – certainly at the cost of thousands of lives – to allow the US and the authors of the war to avoid the reputational splatter when the end came. That is an understandable but certainly ignoble aspiration.

This is the “decent interval” that Richard Nixon and Henry Kissinger engineered for the fall of South Vietnam – basically an effort to game the 1972 election with “peace with honor” and leave others to pick up the peaces. As it happens, when Saigon finally fell in April 1975 Nixon was just months into his forced retirement and California exile, having been forced into the first presidential resignation in history. Life’s a bitch.

Get TPM in your inbox, twice weekly.

Your subscription could not be saved. Please try again.

Your subscription has been successful.

So the government fell – clearly quicker than the White House or the Pentagon anticipated. As things spun out of control in the 2nd week of August, the US redeployed first a thousand and eventually more than six thousand troops to Kabul. As it unfolded this was portrayed as evidence of disorganization and failed planning. You just pulled out! Now you’re going back in! WTF!?!?! In fact, it was what was required to manage the evacuation that unfolded over the following two weeks, a deployment in sufficient force to allow the US to manage a rushed but orderly evacuation on its own terms. Thirteen Americans troops were killed in that effort. But given the vulnerability they were exposed to – essentially searching and vetting Afghan civilians at close quarters in a situation in which literally anyone could approach the US checkpoints – the risk to those Marines at the gates was vast.

So I would ask again, what exactly was the great failure of execution? There have been a few criticisms on operational grounds. But they seem weak at best. One we’ve heard a lot is that the US should have held Bagram Airport rather than the civilian airport in Kabul. I’m not a military planner. But on its face if the main mission is evacuation and the people are in Kabul it seems clear that the airport near the people is the better option. Otherwise, you’re in a sort of Mad Max type situation managing highly vulnerable convoys on a forty mile drive from Kabul to the heavily defended military airbase. I’ll leave that question to others. But again, what was the failure of execution? It really seems that the “failure,” the “there had to be a better way” argument, is that the government fell. Is that a failure of execution? Not really. It was an inherent risk of withdrawal. Indeed, by any candid evaluation it was an inevitable result of withdrawal. The question was just how long it would take to happen. It happened faster than anyone seems to have anticipated. And the US reacted quickly with a contingency plan to manage an evacuation which was actually quite successful.

Let us remember that two weekends ago a crush of prominent commentators and reporters were stating as fact that the White House had been caught flatfooted and abandoned everyone who had worked for the US during its war in Afghanistan. They declared the evacuation a catastrophic failure and shameful betrayal when it was actually only starting and then in many cases took credit for the evacuation after it happened, on the premise that it was only their cries of betrayal that made it happen. This is a bracingly self-serving and sloppy logic. One of the worst offenders on this front, ABC’s Martha Raddatz, was forced to move the goalposts yet again yesterday: now judging the evacuation a “success,” she claimed Biden was “conflating the withdrawal with the evacuation … they did not realize the Taliban would take over so quickly.”

And there you have it. The problem was the withdrawal itself. That is by definition not a problem of execution but one of policy. Raddatz’s argument appears to be that the US should have known the government would fall in a matter of days and that if it knew this it should have stayed long to ensure it would last at least weeks or months if not years. This is arguing what amounts to a distinction without a difference verging on the logic of perpetual war and occupation which is what kept the US in Afghanistan for twenty years in the first place. It’s the final redoubt of a bad and for many deeply dishonest argument. Raddatz is like a desperate evacuee clinging to the skids of a departing helicopter as her preferred storyline collapses around her. Messy, indeed.

The one point she and others are right about is that Joe Biden owns the withdrawal. He and his White House team are now taking credit for that. They argue (correctly) that the American public wanted to leave and that he stuck to his decision even in the face of a storm of criticism. Three Presidents understood the futility of the mission. Only one had the determination to end it even at the cost of real political damage to himself. That means he has to own the reality of withdrawal, the acceptance that the mission, as it expanded in the years after 2001, did not work. He has to accept the reality of a Taliban government. He has to accept the reality and images of terrified refugees, masses looking to escape.

But as many have argued this was a reality baked into the futility and failure of the mission itself. There was no pretty exit. That is what kept the US there for two decades. As has been the case for weeks, this is the crux of the ‘there had to be a better way’ crowd’s argument: wanting out of a failed endeavor but unwilling to stomach let alone embrace the reality of that failure and eager to pass that messiness off on someone else.

Read the whole story
duerig
51 days ago
reply
The more I read about this, the more I realize that the US-backed government falling 'sooner than expected' was baked into the cake. Everybody believed that it was going to fall. As soon as that happened, the fall was going to be instant. Nobody wants to be the last person fighting for a cause that is abandoned by its own leaders and its allies. This isn't anything special about Afghanistan or the Taliban. This is trivially true of every civil war that every happens. This evacuation seems like it was the best of all possible worlds given the 20 years of history that led up to it.
lamontcg
51 days ago
And the "middle managers" in Afghanistan at State/DoD were likely people too stupid to trasfer out of a losing position with a few hard working altruists without enough support to be effective and holding back the tide
Share this story
Delete
1 public comment
acdha
51 days ago
reply
“Raddatz is like a desperate evacuee clinging to the skids of a departing helicopter as her preferred storyline collapses around her.”
Washington, DC

Google’s New AI Photo Upscaling Tech is Jaw-Dropping

2 Comments and 3 Shares

Photo enhancing in movies and TV shows is often ridiculed for being unbelievable, but research in real photo enhancing is actually creeping more and more into the realm of science fiction. Just take a look at Google’s latest AI photo upscaling tech.

In a post titled “High Fidelity Image Generation Using Diffusion Models” published on the Google AI Blog (and spotted by DPR), Google researchers in the company’s Brain Team share about new breakthroughs they’ve made in image super-resolution.

In image super-resolution, a machine learning model is trained to turn a low-res photo into a detailed high-res photo, and potential applications of this range from restoring old family photos to improving medical imaging.

Google has been exploring a concept called “diffusion models,” which was first proposed in 2015 but which has, up until recently, taken a backseat to a family of deep learning methods called “deep generative models.” The company has found that its results with this new approach beat out existing technologies when humans are asked to judge.

The first approach is called SR3, or Super-Resolution via Repeated Refinement. Here’s the technical explanation:

“SR3 is a super-resolution diffusion model that takes as input a low-resolution image, and builds a corresponding high resolution image from pure noise,” Google writes. “The model is trained on an image corruption process in which noise is progressively added to a high-resolution image until only pure noise remains.

“It then learns to reverse this process, beginning from pure noise and progressively removing noise to reach a target distribution through the guidance of the input low-resolution image.”

Before (left) and after (right) upscaling with SR3.

SR3 has been found to work well on upscaling portraits and natural images. When used to do 8x upscaling on faces, it has a “confusion rate” of nearly 50% while existing methods only go up to 34%, suggesting that the results are indeed photo-realistic.

Before (left) and after (right) upscaling with SR3.
Before (left) and after (right) upscaling with SR3.

Here are other portraits upscaled from low-resolution originals:

A selection of portraits upscaled from low-res originals by AI.
A selection of portraits upscaled from low-res originals by AI.

Once Google saw how effective SR3 was in upscaling photos, the company went a step further with a second approach called CDM, a class-conditional diffusion model.

“CDM is a class-conditional diffusion model trained on ImageNet data to generate high-resolution natural images,” Google writes. “Since ImageNet is a difficult, high-entropy dataset, we built CDM as a cascade of multiple diffusion models. This cascade approach involves chaining together multiple generative models over several spatial resolutions: one diffusion model that generates data at a low resolution, followed by a sequence of SR3 super-resolution diffusion models that gradually increase the resolution of the generated image to the highest resolution.”

Google has published a set of examples showing low-resolution photos upscaled in a cascade. A 32×32 photo can be enhanced to 64×64 and then 256×256. A 64×64 photo can be upscaled to 256×256 and then 1024×1024.

As you can see, the results are impressive and the final photos, despite having some errors (such as gaps in the frames of glasses), would likely pass as actual original photographs for most viewers at first glance.

“With SR3 and CDM, we have pushed the performance of diffusion models to state-of-the-art on super-resolution and class-conditional ImageNet generation benchmarks,” Google researchers write. “We are excited to further test the limits of diffusion models for a wide variety of generative modeling problems.”

Read the whole story
duerig
53 days ago
reply
This looks really neat and is potentially a great way to 'enhance' entertainment. My real worry about stuff like this is that the 'enhancement' is simply a plausible filling in of the picture. But since it is magical AI algorithms (tm), people will believe that it represents truth. Will we see people try to 'enhance' grainy security footage and then finger a suspect because the higher-resolution photo looks like them? This seems like a great way to give ourselves false certainty about things we need to remain skeptical of.
ChrisDL
53 days ago
Definitely. I really wish they had also added what the actual original looked like so you could compare it with the upscaled version.
Share this story
Delete
1 public comment
m
52 days ago
reply
I

Scientists Are Working To Make Sense Of 'Long COVID' And Its 203 Possible Symptoms

2 Comments

It's not clear why some people who get COVID-19 are plagued with symptoms for many months after being infected, but scientists are investigating what's behind these "long haul" cases.

Read the whole story
duerig
58 days ago
reply
And if the flu was much more deadly, and if the flu caused permanent brain damage, and if the flu was much more contagious... Its basically the flu, though. :/
Share this story
Delete
1 public comment
JayM
58 days ago
reply
Just a bad flu. I mean… I guess if some parts of the flu in some people lasted for months (maybe years, maybe a lifetime)…
Atlanta, GA

Steel made without fossil fuels delivered for 1st time

1 Comment and 2 Shares
HYBRIT fossil fuel free steel

Swedish green steel venture HYBRIT says it has made the world's first customer delivery of steel produced without using coal as it looks to revolutionize an industry that accounts for around eight per cent of global greenhouse gas emissions.

Read the whole story
duerig
64 days ago
reply
This could be the start of a really big change for the world. I never would have guessed that coke-less steel was even close to a reality.
Share this story
Delete
Next Page of Stories