It's honestly a super subtle difference and only structure heads like me care about it. It matters on the edges. My advisor gave sage advice that I think more people should take to heart. To paraphrase: every experiment has limitations and assumptions baked in assumptions. That doesn't mean their results are invalid/irrelevant, but you need to know what they are so you know when they are violated or don't apply.

It's a spin on the "all models are wrong, but not all models are useful

/pedagogical soapbox

It's a great use of AI / machine learning tech. Incredible. Turns out biology reuses structure a LOT. Structure is function in biology, and there are a LOT of shared, essential functions in Biology. Their models are actually incredibly accurate at predicting the individual atom placement for side chains (the bit of an amino acid that makes it unique from other amino acids). Side chains do chemistry for proteins, so this is highly salient for research broadly. It's just far from being a "solved problem" like they would have you believe.

The main thing to keep in mind, is that alphafold is not predicting structures based on first principles (that is to say, based on the underlying physics and laws of nature). It uses sequence similarity between proteins with solved structures to make probable guesses as to the structure and how it folds. Solely based off current experimental data-driven structure models.

This works surprisingly well even for proteins that don't have much actually in common with the amino acid sequence of the protein. But because structure is function, we can still trace and track the divergences in sequence over time while still being confident the overall shape is the same.

But for things that there are not enough sequence diverse examples of, or for things that there are no examples, alphafold regularly just spits them out like a literal ball of spaghetti - because the assumptions it relies on are invalid. There's not enough statistical evidence - for those examples.

I don't have a handy reference offhand, it's been a few years, but there is this image from their blog posts ( ref ) that shows whe I'm talking about. My understanding is that AlphaFold is an incredibly accurate and effective homology modeling strategy (how can we model structures we don't have data for based off similarity to structures we do have).

Disclaimer: I am not an expert in structural prediction, homology models, machine learning, structural determination via crystallography or cryo-EM. I'm more experienced in consuming them for understanding the structure:function relationship.

Not knowing your expertise, I'll risk making a fool of myself.

But first, for everyone's sake: ALPHAFOLD ISNT AN LLM

That isn't what alphafold is. Under no circumstances is it anywhere near close to a "database of every way ... proteins can fold." At all. And to construe it as anything other than an attempt at a way to predict or simulate the potential three dimensional structure a protein can adopt will mislead people.

A protein does not have one shape. It has multiple. They're dynamic. They change shape. It's how proteins represent information. I'd argue that the ability of proteins to change their shape is one of their most important properties. (Analogy: play a song on an instrument with one note. Hard-mode: consider silence a note)

Everything in Alphafold is a GUESSED MODEL and not reality. Crystal structures and cryo-EM structures are also MODELS. But they're based on empirical evidence.

Alphafold is based on statistical evidence. It is evidence, but it is weaker. If we don't have an example of how a protein might fold in the structural database, alphafold will struggle to predict the structure. At least not without it sharing some type of sequence similarity.

I see this in the AI drug companies and how they just treat predictive models the same as 2 angstrom crystal structures and it pisses me off.

Here is how it concerns you:

By restricting what people can and can't do (by removing features you don't use), they slowly restrict our collective ability to independently verify their claims or to independently serve our own needs - making us more reliant on them.

It's small and minor, but it's just like with the DMCA and the new ID laws. It's about slowly erecting barriers that impede our ability to act collectively or anonymously under the guise of it being "for our own good."

Yes. People need to get out and fucking vote. Anything less than 80% participation in a presidential election is shameful.

But you need to do some background I think. California 22 might be "only" 3k, but it was 53%-47% in 2024. It's also voted Republican since 2002 (1) The real headfuck about CA-22 is how it's 73% Hispanic and reliably red.

Same shit with CA-3. "Only 24,000 votes" - they won that district by ten bloody points 55%-45%. There were 400,000 votes in the district. Thought it's a little more purple. (2)

Also, your numbers are off compared to wiki. Not sure where the discrepancy is but CA-22 and CA-3 were lost by 11,000 and 46,000 respectively.

A good journalist does not just write for themselves, they must also write for the audience as the audience. They are the readers proxy and ask questions on behalf of the reader.

Imagine a generic shmuck that has no idea ctrl +c and ctrl +v are hot keys.

Now. Do you think they have any concept of what type of duplication tools exist? Do you think they'd just want to use this "super cool everything" software that AI is billed as? After all, it's supposed to be smart right? They say it's the next best thing and it's almost like magic.

Look, I'm not saying it's a great article. But given all the bullshit hype regular people hear about AI. Is it really unreasonable to think "copilot, help me find duplicated files in my one drive" would be something good old Billy would try?

Maybe it's not a great article because there are better ways to de-duplicate photos. But that isn't the fucking point of the article. The article is "look at how AI still fucking fails at basic shit we expect it to be good at."

And for that, I thank the author. We need way more of that.

It's both. It's insane to me someone can watch animals instinctively display insanely complex behaviors untaught (e.g. herding by australian shepards) and the scientific research to reduce aggression in a related species before coming to the conclusion that there is no way whatsoever that nature is a significant component. Oh, and just completely ignore breeds bred for traits and behaviors seemed desirable for every domesticated animal.

Nature has no place at all it's only nurture. Sure.

Because it will not be enough.

Because they will come back and say "look at this loophole"

"Think of the children" you'll all say as you agree to give your government authority to determine what information you can or cannot access as "age appropriate" completely ignorant of what you're handing over.

This would be fine if it was just for you, but you're trying to give my control over my system and what I can access away from me because you're too short-sighted to see what comes after volunteer age reporting. And when that still doesn't save the children, which it won't, because it is NEVER ABOUT THE FUCKING CHILDREN ITS ALWAYS ABOUT CONTROL, you'll tell me again that it's just another little minor infraction. It's just a little bit more than volunteer reporting.

Afterall, won't someone please think of the children?!

I would propose a more generous starting position than "you're wrong." Maybe "our understanding is incomplete."

There are many stories that could be a long the vein of one I saw personally: a night-shift leaning researcher can't get a protocol to work after being trained on it. They go back to the post-doc that trained them to troubleshoot - works flawlessly. When they do it by themselves (typically at night) it fails again.

After much agony and self-blame, turns out the enzyme they were studying was regulated by the organism's circadian cycle. They protein they were studying was off at night and no one knew.

Contradiction in science frequently masks a deeper biological truth.

Jesus titty-fucking christ, what a racist take.

While that's correct, I think you're ignoring how much sway a figurehead can hold.

Also it's important to keep in mind, that the childfucker-in-chief may just classify the committee as political appointees and fire them (See Project 2025 and Schedule F). Legality doesn't matter.

He wrote an executive order clearly in violation of the Constitution to remove birthright citizenship, and the Supreme Court said "the real unconstitutionality is nationwide blocks."

I dunno. I'll probably get hate for this, but it's not ready. It's better. But Linux isn't a good replacement for Windows yet. I love Linux. Love the customization, the *NIX filesystem makes sense, and it's beautiful. Also no ads in my start menu!

I want to use Linux regularly, and I tried last week. It failed. Kind of miserably.

I need to pick a distro. Mint and Pop_OS were consensus recommendations.

Try mint: Installing dual boot alongside windows was beautiful. But no internet connection, says cable is unplugged (it's not). Realize I downloaded an earlier version (20). Get the most recent version, and problem resolved. It's kind of odd to me that even a pretty recent version wouldn't support my adapter, but whatever. I tried to update and install Nvidia drivers: update fails because dependencies were not installed. Okay.... Why not prompt me to install them? Why make me apt-get all the dependencies by hand? I don't expect handholding, but some things should be. If I NEED something as a pre-req for what I'm trying to do, queue it up!

Fuck it. Let me try Pop_OS, instead - that has some gaming chops, right? Dual boot was more challenging to stand up, but it all worked. Nice. Fire up game: get ~20 fps drop compared to windows (108 from 130) with the same settings. I don't want to troubleshoot the performance hit. It should just work. I want a tool not a project.

Never mind if you want HDR support. That seems to vary by distro. Variable refresh rate also seemed to be spotty from what I read in gaming distro recommendations. ALSO, do you need UEFI support? RIP. Enjoy toggling that on and off when you have to jump back and forth between Windows and Linux. Nvidia driver support I chalk up to those arseholes only now starting to open source some things.

And I don't care that you were able to run everything fine. You had a flawless experience: great. Love that for you. I didn't. I'm not a computer novice - I know to Google shit and how to implement it. I remember trying to fuck around with Ubuntu back in 2002.

I'm gonna continue trying to stand up Linux for everyday use because I love Linux and I want to use it, but it's pretty clear that even as someone that wants to use Linux. I've been trying to switch to Linux every few years for decades. It's still far short of being ready for average users.

It's not even strictly what capitalism is about. It's some stupid bullshit interpretation that came out of the University of Chicago economics department.

Seriously, go look at Adam Smiths wealth of nations. The only mention of the "invisible hand" is so different from what is taught in economics now

… by directing that industry in such a manner as its produce may be of the greatest value, he intends only his own gain, and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention.

It's not that we need no regulation. It's just trying to say that if we set things up correctly, we don't need to worry about people pursuing SOLELY their own personal gain. Because the market seeks out the "greatest value", which is not just about money. It's also the value to society as a whole.

Instead, we got the fucking bullshit from Chicago saying that the only / best way of measuring value is by profit.

You all are missing the forest for the trees! LLMs are just like us in how we think! We're all just probability generators! No, they can't think or reason beyond known data sets. Yes, they fail at extrapolating information which is the basic component of reasoning. But you guys don't get it! They're just like us and smart!

I'll give this one go just on the off chance you're being genuine - since you seem to have some sense of rationality to your position.

You're wrongly underweighting the damage that will get done while the courts correct things. Yes, eventually the courts would restore broadcasting licenses. But there's appeal after appeal after appeal. That could kill the stations even if ultimately they'd win. And don't say that the licenses couldn't be revoked in the first place. They did it to fucking science funding. Funding mandated by Congress just stopped, and Congress did nothing. Sure it's coming back now. Kind of.

We know how this plays out already. And we can't even count on the courts. The supreme Court has shown a total disregard for their duty with some of their rulings.

Like the executive order on birthright citizenship. Before the supreme Court the government didn't even argue the legality. They said nationwide blocking orders were the real problem. The Court agreed.

Or the use of race or language as a cause for ice detention. Blatantly against the forum amendment and the Supreme Court said it's fine. They're snatching US citizens for fucks sake.

This is all easily searchable. If you really care about being rational and reasoned, you'll put in the effort. This isn't people desperately trying to be wrong. These are people that have seen what is likely to happen. They've updated their priors more accurately than you.

Every fucking year hundreds of thousands, if not millions, of Americans pay a tax on an illiquid asset: the property tax.

Don't fucking tell me "oh it's too difficult, there's too many effects." Tax. Them.

Progressive tax rate on long term and short term capital gains. 95 % upper limit for income tax. Kneecap their bullshit "buy, borrow, die" cycle.

Get back to the fucking WW2 tax levels, back when we had a middle class, and then we can maybe finally have a balanced budget, at least.

D E M O C R A T singular, one. Not democrats. For fucks sake it's on the bloody title!

Why are people so willfully ignorant?

US Rep. Zoe Lofgren (D-Calif.) today proposed ...

There's a lot of people here unable to recognize how important attention span is for deep, meaningful learning.

Way too many people saying "ATTENTION SPANS NOT INTELLIGENCE DUHMASSES" and failing short because they don't have the attention span to think of implications of decreased attention span.

Long term memory is built on repetition and attention. It is entirely plausible that short form media makes you stupider - by impacting your ability for deep learning

It's not Luddism to recognize that foundational knowledge is essential to effectively utilizing tools in every industry, and jumping ahead to just using the tool is not good for the individual or the group.

Your example is iconic. Do you think the average middle schoolers to college students that are using AI understand anything about self hosting, token limits, and optimizing things by banning keywords? Let alone how prone to just making shit up models are - because they were designed to! I STILL get enterprise chatgpt referencing scientific papers that don't exist. I wonder how many students are paying for premium models. Probably only the rich ones.

don't upload it to the internet!

or use a smart phone

or corporate searches that track you

or go to any website with ads - they track you

hell don't even search the internet! your ISP tracks dns requests

or use a modern tv that tracks what is on your screen

or you can do custom phone from - just unlock the bootloader, root it, and install! then just setup pihole/adguard/self-host everything

it's simple, for privacy just go live in a yurt in the woods to not be tracked 24/7

I have the new pixel fold. The first and most important part is to understand how Google views the warranty on the inner screen.

Within two weeks I had what seemed to be hairline fractures on the inner screen and a potential artifact in the fold crease. I wasn't too worried but I was going to have them check on it. Before I could take it in to the Google store, it snapped. The entire screen flooded black within 14 hours when I went back to the store.

I was informed that this wasn't covered under warranty, but they'd make an exception because i literally bought it two weeks prior.

Their stance is that once the screen is that fatally flawed manufacturing defects and misuse damage (i.e. dropping) looks identical. The curvy bendy middle means the outer edges are under more stress and stiffer so when the screen breaks even from defects or creates impact shatter lines.

Based on this alone, I wouldn't recommend anyone get the phone. Not without expecting to have to pay for the insurance plan and to budget for replacing the inner screen at least once. It's significantly heavier and with the fear of breaking, I have a heavy duty dbrand case. So the phone feels like a bloody brick.

That said, I do love it. And I don't know if I'll go back. I don't know if I'll stick to the form factor either. I do a lot of home server shenanigans on it. Home assistant. Control my tv. I live multi-tasking on it when taking notes. I could buy a phone and a tablet for cheaper. But there's something about just having the extra size always handy rather than having to walk around the house with a mini-tablet.

Just be aware that there are huge, glaring downsides to the form factor before you buy in. It is objectively cool and I love unfolding it. It never doesn't feel futuristic.