But I don’t see how any current AI system will have the curiosity and drive. You need to create conditions for meaningful experiments to be done and a mechanistic way of doing this doesn’t work. The most elegant experiments find ways to cut to the core of the emergent behavior, around all the simultaneous changes happening in different levels of the system from which said behavior arises.

You simply cannot see it all. Not every AI can, we are measurement limited. We can track and measure some dynamics, and the art of the experiment lies in thinking through the connections, then picking the right places to look to catch the system in the act. And you better do it so the failure guides the next experiment, and the next and the next, till you finally have sliced through the complexity and found some key nodes that hold the keys.

And usually after all that, you find your key node is regulated by the key nodes of a few other grad students. The complexity here is in part because the entire system is in constant conversation across levels of organization. The causal flow is looped and is also very sensitive to the environment.

AI has a big role to play, but I doubt it’ll parse it without humans at play. It simply cannot invent a slight modification of a microscope and physically connect the new lens element in the right place, calibrate it, all in time for the biology grad student next door to bring some cells to image that are a mix of 40 days old (no contamination! Magic!) vs 30, 20, and so on, to do a very expensive experiment (in terms of labor and time and rats and materials).

So much can go wrong on so many vectors. Which is also what pushes humans to their creative best. And AI will help a lot. But till you have a robot that can drive, but also do mouse dissections and staining and manipulate a microscope with faulty parts to get a clear image, you’re going to need humans to study biology.

When you invent these super robots that somehow slide through all this physical labor without consuming a whole lot of power, we can talk.

Meanwhile, at the cutting edge, where neither procedure nor mass produced machinery exists to do the whole thing, the reality is, we already use a lot of fancy tech in biology, and can use more, and AI can help keep the focus on the actual physical experiments that take so much care and attention.

There’s no well-labeled dataset for all this because we don’t know everything we are looking at.

Take a neuron. You can stain it so you can tell apart an axon and a dendrite. But if you don’t it’s very possible to miss or mislabel the axon, because unlike the textbook neuron, the axon isn’t always obviously oriented in a way that promotes ID.

And that’s just one example. With more complex samples, you also have a lot of edges which AI models don’t segment very well.

That is utterly untrue. It stems from the notion that biology only needs big data for us to solve all its mysteries.

Beyond that, while automated labs for cell biology are plausible and active, and the same fit organoid biology, the same really isn’t true at organism level. Labs already are artificial environments that make organisms have odd phenotypes, and odd interactions which bias the data away from what you’d find in the natural world.

You add in robotic manipulation, and you add a new confound.

Drones in the wild are an obvious counterbalance to this, though even in that scenario, the drone can be a confound unless well hidden.

In the end though, this kind of systemic breakdown of biology, while much needed raw material to improve our theories, won’t solve the whole of biology. Human hands and human minds will be needed for a good long while.

So exactly how can the model believe anything else? It only sees the externalized summation. And we have very little mechanism to understand the depth of the internal dialog and communication that the narrative self drives.

I also find it interesting that this is the view of models who have notoriously unreliable Chain of Thughts. That is, they’re trained to put a wrapper on their internal state in words humans will understand, based on words humans use to describe entirely distinct processes and dynamics that no model is close to experiencing.

It must all seem, from that vantage, utterly fictional.

Comments URL: https://news.ycombinator.com/item?id=50190262

Points: 4

# Comments: 1

We’ll need to build a nose for the right kind of information. Kids growing up in this world already are better at it than older folks who are newly thrown into this environment.

But beyond that I think our agency has already been sapped, as a species. Rebuilding the muscle to exercise our agency as free citizens using the information at our command to make better decisions… we shouldn’t surrender this ability to AI. The best science fiction and science writing already gives us plenty of reason to avoid this. We don’t need to get to the worst of it, and have all out war to correct it.

And that does mean investing heavily in human education. Getting better at teaching things to adults. Making insight easier to gain at every strata of society so true freedom of information when it comes to public works and governance is achieved, which makes the people the watchers and owners of government.

No easy thing to do. Absolutely doable, technologically speaking and in terms of our current biology.

What we have is years of thinking we’re not capable enough. Or cooperative enough. Or rational enough. We can be. We just need the right context, a full belly, and practice in using our voices and shaping our futures actively.

The early evidence isn’t good, but I do believe with deliberate and thoughtful training our brains our capable of much more than we currently give them credit for.

I suspect some will. But to find out, we will have to have a trial, so no one is buying it as truth without checking and handing out AI medicine candy. That too, though, will likely change if the peptide craze is anything to go by.

This doesn’t mean they fold as a company, or give up on AI, or stop working. But what exactly is the explanation for continuing to train ever larger frontier models in the current context?

The AI company leaders giving these warnings were giving the exact same warnings before they had AI companies - and in many cases, before they worked in the industry, or before the industry existed.

And that doesn’t stand the sniff test, because:

1. There’s a strong linguistic silos around these AI leaders and many of their employees. Assumptions couched in language that should be rigorously questioned is instead sometimes axiomatic. 2. It’s hard to understand why, if you had the same exact fear before starting the AI company, you started it in the first place, let alone continue even as you bemoan these fears. 3. Why, if this path is so dangerous, these leaders don’t consider alternate paths to building AI. Why are all their investments in this one way to do intelligence? Surely we will benefit from having multiple routes to choose from?

The absence of choice and the piling of inevitabilities is why these warnings don’t add up. That the same drum has been banged by the same people for a few decades without input from other fields and sources is what makes all this seem religious prophet adjacent rather than technical assessments by thoughtful engineers and scientists.

1. Find patient organizations/support groups for your cancer. They often have excellent collections of info, from some of the sources covered in the article, and can help with finding the right doctor, or the right place to get a second opinion, or financial assistance. Some also list recruiting clinical trials.

2. Try and find local support groups. Other patients are often going to give you way better ideas and suggestions that you never thought of. It also helps enormously to have other folks going through similar journeys to talk to.

3. If you can afford it, Sloan Kettering does a remote consultation, where your primary oncologist can also be present, and they review all tests, suggest new ones, go over the treatment plan, etc. They do this in some international markets too. Pricey AF, but worth it for complex cases.

1. The transgenesis was successful but initially at least induced behavioral changes. Over multiple generations these hopefully attenuated, but that’s the price of a model system: you poke the fundamentals and sometimes it can change what matters to what you’re studying.

2. I wonder how much being in the wild affects the quality and variety of camouflage, that will be sanded down in lab grown cuttlefish. They have fewer textures they experience, and rarely do their camo in the lab when under true predatory threat. For a question like: does the camouflage adjust when the cuttlefish is tracking the movement of a nearby predator, and what’s the neurological basis of that, are going to necessarily remain hard to answer in a lab.

Still, fascinating work that I hope teaches us a lot about these amazing creatures and ourselves.

Cells are autopoetic. They write their own experience.