People are just going to carry a patch of fake grass around with them to bypass the checks. So I suspect the reviews will be astroturfed.
This is my field. There have been three major scandals in the "amyloid" field -- Sylvain Lesne, Eliezer Masliah, and lecanemab trials not informing patients they had the APOE4 variant, which is associated with cerebral amyloid angiopathy, a vascular condition that the same scientists who led those trials previously noted was correlated with cerebral hemorrhage side effects. Sylvain Lesne produced shitty research that was not that highly cited, below is my take on the field and my concerns regarding journalists poorly communicating the science in this story. My perspective on why we are behind other fields is in paragraph five. The next three paragraphs are just context as to why I think there is no singular “amyloid hypothesis,” and why this kind of journalism threatens our field despite a desperate need for dealing with fraud, too. I realize this is long, I am not a journalist, I am not a good communicator, I am a scientist. If anyone has advice, please share it with me, likewise with questions.
Not being clear about "amyloid" nomenclature threatens to throw the baby out with the bathwater, which will stall an Alzheimer's cure even more. Most proteins are "globular," think kind of round balls of scrunched up string, arranged in alpha-helices and some beta-sheets. "Amyloids" are spine-like fibrillar protein aggregates, where each vertebrate is a flattened version of a protein folded in beta-strands. The vertebrate is created by these beta-strands stacking into beta-sheets. Google "amyloid fiber" versus "globular protein" to see what I mean by this description.
"Amyloid BETA" is a usually disordered protein which can aggregate into ONE of the amyloid fibrils seen in Alzheimer's and other dementia patients' brains. Tau can form amyloid fibers, so can TDP-43, TMEM, alpha-synuclein, etc. This is a good link -- https://people.mbi.ucla.edu/sawaya/amyloidatlas/ -- if you want to see the cross-sections of all of the amyloid fiber structures the field has solved with useful annotations. One "amyloid hypothesis," for example, is that TAU hyperphosphorylation (the addition of a lot of very negatively charged post-translational modifications, think chemical ornaments you can add to a tree) leads to tau amyloid fibers, which then lead to amyloid-beta amyloid fibers. There is lots of speculation about the mechanism, whether amyloid fibers can also have enzymatic function that lead to metabolic dysregulation, whether certain amyloid fibers are actually functional and exist in everyone but that certain types or a certain amount is associated with disease, the catecholamine hypothesis is something that can’t be discounted, maybe amyloid fibers are just a downstream effect of a true ultimate cause (in which case, amyloid fibers are still an important clue, lots of people are doing experiments now where they see how they can get certain amyloid fibers in vitro using co-factors which may be one step back to the root cause).
Another "amyloid hypothesis" is that amyloid beta OLIGOMERs, some kind of non-fibrillar aggregate that we don't know the structure of but that we know contains proteins that usually also make amyloid fibers , causes Alzheimer's. This is what "amyloid beta *56" is, by the way, an oligomer, and what Lesne’s work argued. We find oligomers to be extremely hard to work with and I could write a paragraph or two about why, but the fact that this is true makes OLIGOMER research, some of which is probably legitimate, an easy target for fraud. When molecules are well-known to be extremely difficult to work with, if you try to replicate someone's experiments and you can't, it could be because the molecules are extremely difficult to work with or because the authors whose experiments you tried to replicate committed fraud. It’s easy to think “well, it must be me,” which is how people get away with it for so long.
So, why are we so far behind? Something important to note is that amyloid fiber structure on an atomic level has only recently been cracked. Consider that we've known the structure of DNA since the 50s, but we didn't know the structure of the MOST common amyloid fiber specific to Alzheimer's patients until 20 freaking 16 -- https://pubmed.ncbi.nlm.nih.gov/28678775/. The reason for this is that the method used to solve the structure of most proteins is x-ray crystallography, but nobody has ever successfully crystallized amyloid fibers except for very small fragments of them -- https://pubmed.ncbi.nlm.nih.gov/15944695/. People think this makes physical sense for reasons related to crystal symmetry. The 2016 structure was solved using cryo-EM, which is a relatively recent development and only won the Nobel in 2017. Prior, it was derided as “blobology” because you would get very coarse structures from cryo-EM -- https://pmc.ncbi.nlm.nih.gov/articles/PMC2726924/. So, the field had to wait for cryo-EM to improve, a big part of which was waiting for better computer vision algorithms (i.e. we use YOLO and various CNN-based algorithms, too, we had to wait for that shit just like Waymo), among other things.
Why is structure so important? Drugs have to physically interact with some kind of protein target. So, you should probably know the atomic model of the target so you can make use of computational modeling techniques that can help you figure out what binds to proteins. Alternatively, if you want to do a ton of biochemical screens, i.e. make a bunch of the target protein, treat it with various compounds, and see what sticks, you have to make sure you are correctly making the target protein. In other words, you don’t know you’ve successfully reproduced the target if you don’t know what the target is. And the exact fiber structure matters, since different structures appear to be correlated with different diseases. Then, you have to figure out a method to make that protein, which people also have done recently for various types. The alternative is waiting for brains from brain banks, which is very slow and doesn’t provide a lot of material. Keep in mind this also means you can’t verify your mouse models have the same structures, too.
I do not think these scandals significantly stalled an Alzheimer's cure. It's a genuinely tough field -- you can't do brain biopsies, amyloids are a tricky protein to work with, we didn’t know the structure until 9 years ago. I think there are a lot of people who feel like they were locked out of funding opportunities because of the focus on amyloid-beta. Maybe this is true, I don’t know. The lab I’m in has worked on tau for decades.
Another point -- neurodegeneration starts far before symptoms show up. A lot of the recent drugs were designed for what is basically metastatic cancer. Learning more about earlier stages of neurodegeneration, which we can do with PET-ligands designed to bind to fibers, the recent p-tau blood test, etc. is necessary to uh... treat stage 1 or 2 cancer equivalent.
Finally, I have to get political here, considering recent events. I’m not a professional communicator. Most scientists aren’t. It is things like this that are the reason we are being threatened with funding cuts. Who will read nine paragraphs that I think are necessary context for all of this? But it is much easier to read an article like this, where the same point is repeated multiple times, with some “they said this,” “they said that,” etc. than to understand even a small portion of a field. More people will do the former, and then apparently call for executions without any way to judge who will be executed. And I sit down to write code for my experiments, click on one link, and see what I perceive as harmful information, and there goes apparently half an hour. I can either let this kind of stuff lead to my funding being cut, or reply to it and slow down my research.
In a sense, Adrian Thompson kicked this off in the 90's when he applied an evolutionary algorithm to FPGA hardware. Using a "survival of the fittest" approach, he taught a board to discern the difference between a 1kHz and 10KHz tone.
The final generation of the circuit was more compact than anything a human engineer would ever come up with (reducible to a mere 37 logic gates), and utilized all kinds of physical nuances specific to the chip it evolved on - including feedback loops, EMI effects between unconnected logic units, and (if I recall) operating transistors outside their saturation region.
Article: https://www.damninteresting.com/on-the-origin-of-circuits/
Paper: https://www.researchgate.net/publication/2737441_An_Evolved_...
Reddit: https://www.reddit.com/r/MachineLearning/comments/2t5ozk/wha...
It was never safe for any government to move any secrets to any cloud. The fact that the US government is okay with doing this with its own secrets surprises me to this day. You have no secrets from the person who owns your hardware.
It's pretty telling that schools like Penn don't cut their administrators, but instead they cut their admissions.
"Between 1976 and 2018, full-time administrators and other professionals employed by those institutions increased by 164% and 452%, respectively. Meanwhile, the number of full-time faculty employed at colleges and universities in the U.S. increased by only 92%, marginally outpacing student enrollment which grew by 78%.
When we look at individual schools the numbers are just as striking. A recent report I authored found that on average, the top 50 schools have 1 faculty per 11 students whereas the same institutions have 1 non-faculty employee per 4 students. Put another way, there are now 3 times as many administrators and other professionals (not including university hospitals staff), as there are faculty (on a per student basis) at the leading schools in country."
https://www.forbes.com/sites/paulweinstein/2023/08/28/admini...
For those who want to have a substantive discussion on the federal budget and believe these cuts are justified, I have a few questions (putting aside questions of constitutionality for this thread):
1. There are claims that federal spending is out of control. How do you square that with the fact that spending as a percentage of GDP is only slightly elevated compared to the historical average going back to at least the 1970s, with the main deviation in the past few years coming from the after-effects of the pandemic? [1]
2. Federal spending largely falls into a few categories: taking care of the elderly (36%), defense and veterans (20%), taking care of the poor or disabled (22%), and interest on existing debt (13%). [2] This adds up to 91% of the budget. The US population is aging, which means that 36% slice is going to naturally grow. What do you think should be cut, and how?
3. The US pays far more for health care (28% of the budget if you include Medicare) and with worse outcomes on average. Why shouldn't the health insurance industry be the first item on the chopping block?
4. Corporate tax receipts have been steadily falling as a percentage of GDP. [3] Why shouldn't corporations (that benefit from a healthy and educated workforce, a safe and secure environment, a working transportation system, etc.) be paying their fair share to keep the national debt in check?
[1] https://fred.stlouisfed.org/series/FYONGDA188S
[2] https://fiscaldata.treasury.gov/americas-finance-guide/feder...
[3] https://fred.stlouisfed.org/graph/?g=1DUdb
It's ok boys, now you're allowed to pirate all books ever published as long as you don't seed.
My doctor/'s office was just forced to switch to app-instead-of-website for patient interation. Ironically, the old and this new software came from same megacorp I work for (different department, same story). The old web interface was rather good and rather sensible-practical-sane. It was almost as if someone had asked a doctor what he actually needed for his patients, and then understood his answer and also managed to actually implement and deliver it.
The new smartphone app.. Not so much. It looks more like some MBA types managed to solve "what is the cheapest thing we can get away with, lawfully?"As a result, the use case flow (IE the only way you can operate it..) of the app, goes as follows:
1 close the app 2 launch the app and sign in 3 do ONE action 4 enjoy the result of the action 5 repeat from 1..
You might wonder why that is.. Well, that is because your software is not allowed to display any errors - because that might indicate there were bugs.. So instead, whenever an error happens, you just display the '... still loading..' animation... forever. So, technically, there are no errors, no bugs.. "IT IS JUST TAKING TOO LONG TO RESPOND". (spoiler: it will NEVER respond, because hidden behind the screen, is a series of unhandled web api errors..)
But again, as an "internal" employee, I have seen our management claim all this is a huge success (client paid/pays).
Back to using it: When I have to interact with my doctor, I write the texts on my PC, and mail them to myself. Then I cut/paste them from gmail into this wonderful app.
For the hundreds of people reading this article right now - you might be amused to know that you're accessing it from a mac mini on my desk:
https://www.contraption.co/a-mini-data-center/
(The CPU load from this is pretty negligible).
The state of the product's security wasn't unexpected. I was, however, shocked by this part:
> I was willing to overlook:
> The bed costs $2,000
> It won’t function if the internet goes down
> Basic features are behind an additional $19/mo subscription
> The bed’s only controls are via mobile app
The market is clearly too stupid to vote against the rent seeking tech industry. It makes me so sad.
Too right, it was far more problematic than they ever made out.
> The UK government's demand came through a "technical capability notice" under the Investigatory Powers Act (IPA), requiring Apple to create a backdoor that would allow British security officials to access encrypted user data globally. The order would have compromised Apple's Advanced Data Protection feature, which provides end-to-end encryption for iCloud data including Photos, Notes, Messages backups, and device backups.
One scenario would be somebody in an airport and security officials are searching your device under the Counter Terrorism Act (where you don't even have the right to legal advice, or the right to remain silent). You maybe a British person, but you could also be a foreign person moving through the airport. There's no time limit on when you may be searched, so all people who ever travelled through British territory could be searched by officials.
Let that sink in for a moment. We're talking about the largest back door I've ever heard of.
What concerns me more is that Apple is the only company audibly making a stand. I have an Android device beside me that regularly asks me to back my device up to the cloud (and make it difficult to opt out), you think Google didn't already sign up to this? You think Microsoft didn't?
Then think for a moment that most 2FA directly goes via a large tech company or to your mobile. We're just outright handing over the keys to all of our accounts. Your accounts have never been less protected. The battle is being lost for privacy and security.
I was quite confused why the government would create an entire site telling people to stop consuming manga. I mean, I personally don't care for manga/anime stuff, but really?
Turns out it's just a website for Quitman, GA. https://quitmanga.gov
Well, I guess it would be nice if we could have some precedent for the claim that downloading copyright protected information is not in itself a breach of copyright.
It makes sense from the point of view that distribution is the act protected by copyright, not the mere act of copying. If that sounds odd to you then that's probably on purpose, There's been plenty of opportunity to rename copyright to authorrights or something similar, but then people might start wondering how keeping something from public domain for 90 years after the author's death could possibly be about protecting the rights of the author.
Linus response here seems relevant to this context: https://lore.kernel.org/rust-for-linux/CAHk-=wgLbz1Bm8QhmJ4d...
Because this isn't about efficiency.
The people radicalized to these actions are trying to destroy the government. They haven't been especially subtle about it in their writing. The ideal outcome is not the same government outcomes but at 70% of the cost. The ideal outcome is the collapse of the federal bureaucracy so that the lords of capital can scoop things up and create their little kingdoms.
I'm reminded of the story of the Air Force designing cockpits for the "average" pilot, only to find that
> out of 4,063 pilots, not a single airman fit within the average range on all 10 dimensions [0]
Surely, there are so many employees in general that probationary employees aren't needed. And surely, most government employees don't need to purchase things on a daily basis, so we can inhibit their credit card use. And most contracts about XYZ aren't crucial, so we can cancel them.
But, my goodness, there is so much nuance and breadth to the things a government does, let alone the government that is responsible for the largest military and that props up a big part of the world economy, that compounding these rash decisions will have far-reaching and serious blowback. I'm all about efficiency, but why be stupid about it?
The plane that we're all on is being dismantled midair, the engines have been turned off, and we're just gliding now. Gliding or falling, anyway
[0] https://www.thestar.com/news/insight/when-u-s-air-force-disc...
Dumbest change ever. The employee charge cards are there to get rid of red tape. Now workers will have to spend $100 of their time filling out forms to buy $50 worth of office supplies.
Because there are bigger fish to fry, I think people don’t appreciate the sheer cost of the system rebuild that will be required for security reasons later.
There’s absolutely no telling what additional software has been installed alongside existing, or which systems have been modified that would require audit. Purging this will be an absolute fucking nightmare to the American taxpayer.
This may turn into one of the most significant IT incidents in world history.
As someone who once built a large coop [1] then just bought a pre-built shed for the 2nd coop, it's definitely _not_ the _monetary_ solution. You will probably lose money overall for quite some time. I'm still probably underwater.
BUT, there are definite upsides:
- Chickens are very sweet animals, and are quite intelligent. You will grow to love all the silly things they do. You can pet them, they are super soft, and can become quite tame. They can purr.
- I'm told the eggs taste way better, I don't really notice it because I really only eat my own eggs, but perhaps I just got used to them
- It's fantastic to get ~8 free eggs per day (from 13, 3 are not laying this winter)
- Morally/ethically, it seems like the best way to eat eggs if you're caring for them in a loving manner (compare to factory farms)
Consider the downsides:
- You may have to euthanize a chicken, likely by hand (literally) via cervical dislocation. It still ranks among the worst things I've ever had to do in my life. Imagine euthanizing your dog or cat by hand...
- Predators, foxes and hawks, you need defenses
- Veterinary services can be harder to find. Most vets don't want to deal with chickens. However, it also tends to be cheaper than a vet for a dog/cat.
- Your wife may one day want a chicken to live inside the house. You may one day agree to this, and then miss it when the chicken is living outside the house again...
- If you really like eating chicken, you may end up finding it difficult to eat them again in the future after you develop a bond with them.
I think there are more upsides than downsides, but you should think about these downsides before taking the plunge. Don't let it dissuade you. Overall, they have enriched our lives immensely and I would recommend it to others!
1: https://www.anthonycameron.com/projects/cameron-acreage-chic...
In general, people are going to interpret this EO with their own lens. Unsurprisingly, reasonable people may disagree on the merits of the EO as a whole.
However this part of the EO is pretty concerning
> 'The President and the Attorney General, subject to the President’s supervision and control, shall provide authoritative interpretations of law for the executive branch'
and later
> 'No employee of the executive branch acting in their official capacity may advance an interpretation of the law as the position of the United States that contravenes the President or the Attorney General’s opinion on a matter of law'
This can potentially enable an end run around congress and the courts in that the President can easily choose to interpret laws in a manner inconsistent with the intent of congress and courts. Now, we can argue the point and say that presidents have already done so in the past and that congress/courts should have been more specific. However it quickly gets into the issue of the impossibility of congress or the courts anticipating and specifying every detail to avoid a 'hostile' interpretation.
This part of the EO says the president's opinion is the law as far as the executive branch is concerned. Given that the executive branch implements the law, this would imply that the president's interpretation is all that matters. The other two branches have no real role left to play. Given the supreme court's ruling on presidential immunity, this is a dangerous level of power concentration.
Even if you support the current president's goals and objectives, setting up the president as the sole power center is an inherently unstable system. Nothing prevents the next president from having a radically different opinion. There is a very good reason why the founding fathers built in an elaborate system of checks and balances.
This EO, combined with his proclamation that "He who saves the country does not violate any law" paint a very concerning picture. This has, historically speaking, been the language of tyrants. No President is above the law, nor does the President "interpret" the law; that is the domain of the Judiciary.
“Wherever law ends, tyranny begins”
This is the problem with people being OK with executive overreach when "their team" is in power. Eventually, and in fact about 50% of the time - the OTHER team is in power.. and may just push the overreach further.
We should desire that the legislative side actually legislates and each branch of the government holds the other two in check, regardless of partisan control.
Further having our judicial branch become openly partisan while remaining lifetime appointments despite younger appointees with longer lifetimes, is really the finishing touch on this slow rolling disaster.
This new administration lays bare what we've known all along - the legislative gridlock and dysfunction in the house of representatives and senate has made them completely incapable of governing -- the least productive in a generation.
This is opened up an opportunity for a well funded strongman, and the checks and balances that were intended to protect our democracy are now mere suggestions.
> Ai Pin will still allow for offline features like battery level
> https://support.humane.com/hc/en-us/articles/34243204841997-...
The only feature they could think of was “battery level”? That’s hilarious
Most in-house game engines built after a certain point use a non-trivial amount of third party code, including console stuff which is under strict NDA, so it's a huge hassle to open source them. Most iterations of the Source engine use Havok physics for example.
IdTech probably was only open sourced because Carmack pushed for it, but it helps that IdTech of that vintage was all in-house code exclusively targeting the PC. I think the only thing they had to cut out for legal reasons was the patented shadowing algorithm in Doom 3.
This is good for modding but don't be misled, this is the TF2-specific code which sits on top of the still-closed-source Source engine. For example you couldn't port TF2 to a new platform with this, at least not without reimplementing Source or wrangling it into working with one of the leaked Source codebases and dealing with the legal fallout of that.
18F - and its sibling the US Digital Service, before it was eviscerated to become the host for the DOGE parasite - is exactly the sort of thing you would create if you want government to run efficiently. Creating room for well-meaning technical experts to offer their expertise across the government, without being locked into a lifelong public-sector career path, is a fantastic idea. Even at its worst, it's taking work that could otherwise go to beltway-bandit contractors, but in-house, with all the cost and friction reduction that that implies. At its best, it's able to spot cross-agency opportunities that no other agency would see on its own.
If you're still on the fence about giving the current administration the benefit of the doubt, dismantling these agencies based on zero analysis is perhaps the clearest possible evidence that "efficiency" is not the primary goal, or in fact a goal at all.
Tariffs are a tax on the low and middle class. Upper income people will be fine with a 10-25% increase in cost of things they purchase. Regular income taxes are progressive. Tariffs are not.
There is a reason why rich prefer tariffs over a progressive income tax.
Credit to the engineers that built this, but it fills me with rage that Elon has this sort of unchecked power.
How long before this starts getting deployed in safety critical applications or government decision making processes?
With no oversight because Elon seems to have the power to dismiss the people responsible for investigating him.
Anyone not scared by this concentration of power needs to pick up a book.
Clear view from the plane waiting to take off
https://www.reddit.com/r/CatastrophicFailure/comments/1isabt...
Which also brings into mind cockpit distractions but its amazing they caught this on video.
Karpathy gave his initial impression: https://x.com/karpathy/status/1891720635363254772
The pull quote is: The impression overall I got here is that this is somewhere around (OpenAI) o1-pro capability