jackdeansmith

I think it would be entirely reasonable to ban death stars in private hands, the same way I think it's entirely reasonable to ban uranium enrichment in private hands. My point is that it's a question of fact about how risky an AI model can be. I think it's clear that current models are not enriched uranium or death star level, but I don't think there's anything ruling it out in the near future!

>If the biggest danger of LLMs is that they can hack traditional systems

This is doing a lot of lifting. If the biggest danger of LLMs is they could uplift bioweapon development, the situation is different. If the biggest danger of LLMs is they reach capabilities allowing for recursive self improvement, the situation is very different still.

I see a very clear link, very capable AI models are clearly able to find and exploit cyber vulnerabilities at a speed and scale not possible before. In the right hands, that's clearly a weapon

And what are those ingredients for biology, which can be constrained as effectively as uranium enrichment? I'd argue there isn't anything which can easily be restricted or monitored.

What about uranium enrichment?

They can't. You have to explicitly give apps permission to access your photo library.

I think it's a lot harder to have this particular type of communications failure if you're writing on a chalkboard. Imagine trying to write out that whole slide, it would take forever. If you really did have to present that information on a chalkboard, you'd be significantly more likely to write something along the lines of:

"We checked the test data: possible to damage tiles significantly" "Foam that hit wing was way bigger than the tests"

Obviously you can miscommunicate via any medium, but I think the author's point here (which I agree with) is some mediums lend themselves to specific types of miscommunication.

I think this is clearly a joke

>The real tragedy is not that Hims exists, but that it works so perfectly. Every day, thousands of people choose their compounded weight-loss drugs over FDA-approved alternatives, their combination ED pills over established single-ingredient treatments, their algorithmic consultations over actual medical care. They make these choices not because the products are better, but because the entire experience has been optimized to feel more like shopping and less like confronting the mortality and vulnerability that define the human condition.

Strongly disagree with almost everything in this article, but specifically this. The reason people make these choices is not because of slick marketing working against them, it's because the existing process to get medical treatment is paternalistic, hard to navigate and often expensive.

If you want safe and really high quality medical care you should absolutely have a personal physician you have a personal relationship with, who understands your lifestyle, your risk factors for side effects, and your medical needs deeply. How many Americans have that? Maybe a few dozen? The market has responded to just how terrible the existing system is.

I misunderstood your comment then, my assertion is that models which have the same capabilities as current models will be cheaper in the future. I have no doubts that models with more capabilities will be more expensive.

Want to bet? I'll give you 5:1 odds that tokens from a model with some specific benchmark performance (we can sort out the specific benchmarks or basket of benchmarks if you want to bet) will be cheaper two years from now.

In my opinion most Anthropic models are the opposite, scoring well on benchmarks but not always way on top, but quietly excellent when you actually try to use them for stuff.

What would convince you that a system is too dangerous to release?

There was certainly some client side censorship, but also the model weights clearly had some of this baked in. You can try yourself by asking the same queries to r1 hosted by a US host like together.

As someone born in 1997 and working in the software industry, one thing I’ve learned is that the discipline of software engineering is sometimes mostly about understanding the history of a system. Love reading stuff like this because sometimes it really makes some concept click!

Didn't know I wanted it, but I now really want P2P iMessage between mountain summits

3D photo/video features did come to the pro phones flower/bird/food/tree/everything ID is in the new OS, it identified a plant genus in one of my photos yesterday. Pretty cool!

Eric Gilbertson has become more than a legend in the Pacific Northwest mountaineering and peak bagging community, in part by revising the list of 100 highest peaks in the state with manual GPS and theodolite surveys. The writeups are great: https://www.countryhighpoints.com/luna-east-fury-west-fury-t...

EOR oil is lower carbon than non-EOR oil. Greenwashing seems excessive, there is no fraud here, everyone is very upfront about what they are doing.

Carbon in the atmosphere is a stock and flow problem, even if we get to net zero emissions, we have just stopped the problem from getting worse. At that point, we need carbon capture to be scaled up and efficient. I don't think it's premature or greenwashing for technologists to support early efforts to get this tech ready. I don't think many supporters of carbon capture think the goal is to offset all the current emissions. They mostly think of carbon capture as a way to eventually deal with historical emissions.

> The nowcast produced the following estimates: • The biocapacity for the world in 2023 is estimated at 1.5 global hectares per person. • Humanity’s Ecological Footprint is 2.6 global hectares per person, of which 60% is carbon Footprint.

The "of which 60% is carbon footprint" is an important line here. To be very frank, I think the use of some standardized unit like "global hectares" is misleading for a problem like carbon emissions. My distaste for this is that using a land-area-like unit implies that the total resource is fixed. There are many technologies in various stages of development that will increase our ability to sequester carbon, though at large cost, and many that allow generation of clean energy, sidestepping the need for this resource entirely. To this groups credit, they do state clearly that transitioning to clean energy moves their date significantly. A single variable representing "resources" might be a useful rhetorical tool, but you can't use it to understand a problem in a meaningful way.

I'll speculate a bit more that this type of limited-resource framing is intentional and plays into their pro population control stance. I was surprised (and disgusted) to see the population section of their site quote Paul Ehrlich uncritically.

[Paul Ehrlich](https://en.wikipedia.org/wiki/Paul_R._Ehrlich) is notable as author of the 1968 book [The population Bomb](https://en.wikipedia.org/wiki/The_Population_Bomb). This book warned about population growth and opened with the quote:

> The battle to feed all of humanity is over. In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now. At this late date nothing can prevent a substantial increase in the world death rate...

Of course, none of this came true, but Ehrlich stood by the book. I will quote a section of this book that when I first saw quoted, I had to look up to believe it was real:

> While we are working toward setting up a world program of the general sort outlined above, the United States could take effective unilateral action in many cases. A good example of how might have acted can be built around the Chandrasekhar incident I mentioned earlier. When he suggested sterilizing all Indian males with three or more children, we should have applied pressure on the Indian government to go ahead with the plan. We should have volunteered logistic support in the form of helicopters, vehicles, and surgical instruments. We should have sent doctors to aid in the program by setting up centers for training para-medical personnel to do vasectomies. Coercion? Perhaps, but coercion a good cause. I am sometimes astounded at the attitudes of Americans who are horrified at the prospect of our government insisting on population control as the price of food aid. All too often the very same people are fully in support of applying military force against those who disagree with our form of government or our foreign policy. We must be relentless in pushing for population control around the world. I wish I could offer you some sugar coated solutions, but I'm afraid the time for them is long gone. A cancer is an uncontrolled multiplication of cells; the population explosion is an uncontrolled multiplication of people. Treating only the symptoms of cancer may make the victim more comfortable at first, but eventually he dies often horribly. A similar fate awaits a world with a population explosion if only the symptoms are treated. We must shift our efforts from treatment of the symptoms to the cutting out of the cancer. The operation will demand many apparently brutal and heartless decisions. The pain may be intense. But the disease is so far advanced that only with radical surgery does the patient have a chance of survival.

I consider myself to be a committed environmentalist. I sincerely believe that good stewardship of the natural environment is critical to human well-being and that our society needs to get its act together with regards to unsustainable practices. However, as long as so-called respectable environmental groups continue associating with people like this, I will remain skeptical and keep myself at arms length.

My guess is Tim Cook didn't wear one because Apple's media team didn't want every news article to use the screen grab as the header image (which is what happened with Zuc). They want the their manicured marketing material front and center.

By not generating the fixed json structure (brackets, commas, etc...) and skipping the model ahead to the next tokens you actually want to generate, I think

I wouldn't be surprised if by making code 5x cheaper to write, aggregate demand for code goes up by at least 5x, and possibly more. The history of computing is full of technologies that made it cheaper to create software (compilers, debuggers, high level languages, version control, increasingly powerful hardware, widely-available open source frameworks, etc...) and as programmers became more efficient and effective, the scope of problems they've been tasked with has kept growing. I think there is a good chance that LLM coding assistance will be a similar story.

Honestly a very confusing move for Twitter, third party client users are probably the easiest to monetize users that Twitter has. They have already demonstrated that they get enough value out of the service to go out of their way to use a different client, potentially one that they pay for. Make them pay some monthly fee and everyone is happy?

Feels to me like public pressure is on Apple to actually justify their argument that their App Store policies are for the benefit of their customers. If that results in more policies like this that really do improve customer experiences, that's not the worst outcome.

Right, yeah that's more important, there's no way to get a runaway chain reaction with an RTG

Those are RTGs that harvest a couple hundred watts of decay heat for really remote applications. Often used in space probes, etc... Bit of a different safety profile since they're low power and generally don't have any moving parts.

I feel like a good first step would be requiring that any assets over some small cutoff be held in a blind trust. It wouldn't solve all problems, but it sure would solve a lot.

For me, the only thing keeping me from going to 100% electric for my household is that nobody seems to offer an EV with 200 miles of range with AWD and clearance for a rough forest road with potholes. Until then, we're sticking to a short range EV for the city and a crossover for the mountains.