Tag: Artificial Intelligence

Categories
Internet Culture

Day 1691 and Don’t Let History Be Written By Cthulhu Hentai Watchers

I sometimes think my best work is as a reply guy on Twitter. If the future of the Internet is to be made entirely of bots, I shall mourn the creative glory we will all lose.

We will all be missing out on the bizarrely specific interplay between post and reply guy. The signal, node, and repeater that is the feed where humans can still bounce off intuition and humor. They know the algorithm merely as the setting for their gamble.

I’ve been doing what is probably a terribly ineffectual public education campaign on artificial intelligence and its role in improving problems in the here and now.

But I am often cheeky about the ideological disagreements the more practical minded builders have with a pretty standard issue Bay Area sex cult. Having been raised adjacent to a lot of hippies, cultists, religious fundamentalists and any variety of new age woo woo Art Bell types I’m actually generally quite tolerant of our weirdos.

Of course this is a challenge as we build out the most important informational and systemic organizational technology of our species. Math has taken humans pretty far.

And mathematicians are quite often mad so I don’t see how it should be such a worry. Still I myself don’t want to be stuck inside an information hazard where I am tortured for eternity either. And therein lies the tweet

Look I’m just saying it’s my personal opinion, but I think it’s bad that the future of our knowledge graph is being made by those who get off on Lovecraft but like that’s just honestly mostly a personal human alignment thing because I don’t wanna be tentacle tortured in @RokoMijic basilisk

If you are familiar with the odd pockets of specialties you probably know that say furries are particularly good at network security and that transwomen are well represented in artificial intelligence.

Unfortunately other groups represented across the space including hardware, software and the philosophy of the space can be somewhat less wholesome than a fursona or a cute catboy.

The Rationalists have some members with odd extracurriculars in their science fiction. I’ve read some good stuff in the genre. Maybe this interest in fiction is probably how they imagine such horrific futures for us in artificial intelligence. Very improbable ones even.

So it’s on me to joke and show you where the weird can be and remind you not to take this too seriously as the experiment contains all of human knowledge. How we prioritize what is another matter. You might call that alignment.

I’m sure we are all wondering which AI safety researchers want to have sex with the many tentacled Lovecraftian old gods. The number of AI researchers with Cthulhu hentai is non-zero.

So the sake of our right to decentralized compute, and indeed our right to do math and compute, we need as many types people as possible engaging with models as possible.

Find ways to learn this new way of thinking and engaging with information and searching for information with a billion parameters at play. Some corners of the most doom and gloom road have some uncomfortable fixations. And artificial intelligence needs all of us.

Categories
Politics

Day 1689 and Drumhead or Not Rushing To Judge What We Don’t Know

I wonder how much of the moral education of Americans comes thanks to Gene Roddenberry. Star Trek is a very American show. It was pitched to Desilu Studios (owned by Desi Arnaz and Lucille Ball of I Love Lucy fame) as a space western.

The cowboy frontier spirit of Captain Kirk has an honor code based on the worthiness of exploration. When the millenials got a reboot in the Next Generation, Captain Picard added the gravitas of excellence through the pursuit of the truth.

It’s hard to think of a show that embraces an anglosphere manifest destiny with more vigor and it is grounded in the enlightenment values of science and moral philosophy. It’s a very American show.

I was watching a fairly minor TNG episode while eating today The Drumhead. An ambassador starts a witch hunt insinuating treason. Apparent drumhead courts were a military summary judgement court.

A legal drama ensues where Picard quotes the ambassador’s father famous words about the dangers of denying basic rights to even one man in the name of protection of society makes us all less free.

Feeling rather on the nose for the moment right? Picard reminds Worf as it wraps up the witch hunt being proven a lie, that those who cloak their misdeeds with the pretense of serving a greater good are often the most difficult. Spreading fear and mistrust in the name of righteousness must be guarded against.

“Vigilance, Mr. Worf. That is the price we have to continually pay.”

We have a chance at preserving and expanding the enlightenment values that brought our species to this moment, an American moment, by enabling more of us. We can educate ourselves and access the world and the many virtues to which humanity has aspired to by working hard to master them ourselves. This is the empowerment of compute.

We have tools available to us to be vigilant in what is the great good of civilization we wish to protect in an open, decentralized and networked internet.

That we can at scale train a new inference search capacity for the sum of our knowledge and reality is a utopian reality even Gene Roddenberry didn’t dream we’d have.

Don’t rush to summary judgement in fear over the future we can build with protocols and networks that can be encrypted and decentralized. That we can model them with artificial intelligence, and by asking questions well, get back information, code and even machine and biological diagnostics is a huge achievement for our species.

To push the fear of the greater good over its consequence can go from legitimate concerns to a drumhead summary judgement quickly. Risking the freedom to use something that is both just emerging and even in that infancy so promising. That it is helpful in the now seems to showcase both western traditions and enlightenment values.

It’s an American ideal that we must all be free to think and speak the truth, search for the truth and can calculate and provide a mathematical proof of truth that needs no trust because it can be know. We have a right to compute.

Artificial intelligence solves problems now. It’s better than humans now at many types of tasks even as it still a new tool and unrefined and even brittle in its capacities still. It’s took only as good as its craftsman’s understanding of it and we are still learning how to build it.

Artificial intelligence can give us the internet we deserve. I support the right to repair movement because I believe we should know and seek the freedom to own things we have purchased and modify them.

Right now AI helps humans can solve immediate problems from tractor repair to wound care. Maybe machines get better as we get better at using them because we do it with them. We figure things out by building.

Don’t let your worldview be constrained. You can know the freedom of living in a society that values your rights. And the right to compute weaves together many of your most sacred rights. And if we infringe on it for some of us it infringes us on us.

Categories
Biohacking Chronic Disease Medical

Day 1688 and Deciding What To Outsource to Artificial Intelligence for Health and Wellness

One of my Twitter mutuals recently published an artificial intelligence prompt for making an adaptive fitness coach which works inside any of the major large models.

Having recently been saved from going septic and a gnarly deep tissue infection thanks to clever diagnostics from the commercial chatbots I am considering what else they might be able to help me along with in my day to day.

Being failed rather regularly by doctors over a decade of chronic illnesses has made me skeptical of the institutions in American medicine. But having one doctor (a dermatologist) miss a glaringly obvious differential really shook me.

Her dismissal of the details and particulars wasn’t malice, but a function of the systemic inability to put enough attention on the details of the person in front of her. Attention really was all she needed ironically.

I’m sure she didn’t set out to be that kind of doctor, I’d bet she hates that it’s all 90 second visits and Medicare coding and making money for the private equity group who owns the clinic. I feel for her. She surely wants to get back to doctoring.

No one can spot every detail and retain the complexities of every case. Especially one like mine. But a computer has a much better shot at mimicking Dr House than I do at finding a Dr House for myself. And it certainly has a better chance than someone who let the system dominate them into breezing over the details.

So I am using my mutual’s prompt to see if I can outsource a very slow and adaptive return to fitness after my month off from exercise to recover from surgery. I like what I’m seeing from all models that I’ve tried it on but I imagine I’ll have all the same “me” problems with overdoing it and pushing too hard. But who knows, maybe this aspect of wellness is better handled by machine than by me.

Categories
Internet Culture

Day 1464 and Which Information Island?

It would be helpful for most media readers to understand the history of the news business and its relationship to war and finance.

For all the standards and ethics and best practices we expect from professional newsrooms (and they do have conduct standards), the history of media isn’t a clean narrative America went to war thanks to Yellow Journalism in the Spanish American War. If you think the Pulitzer is a badge of merit wait till you learn its history and financing.

You have probably lived through multiple media scandals. Millennials remember the neoconservative “weapons of mass destruction” story thanks to notably terrible editorial decision at the New York Times.

When someone says you are in a media bubble or on an information island, recall that these systems made up of people with varied interests, ambitions and aims.

News

American Surgeon General Calls for Cancer Warning on Alcohol

TechCrunch grudgingly says the AI Bloomers beat the delusional AI Doomers

Essays

An Nvidia researcher Jim Fan tweeted about blog post from the late Felix Hill on 200B weights of responsibility inside the work of building AI. The stress of knowing the billions of weights and bias that make up artificial intelligence that will bend our lives for both good and ill is no small thing. The thread has many beautiful comments.

Vinay the founder of Loom is in the thick of figuring life out after success and it’s brave of him to share.

Applications and Services

AskNews has an interesting set of display options for news types and bias and sentiment I’m considering purchasing and testing.

Categories
Internet Culture Politics

Day 1369 and California SB-1047 Vetoed

Last night I received a push alert and then a flurry of excited text messages and phone calls. California Governor Gavin Newsom vetoed the controversial SB-1047 artificial intelligence bill.

Gavin Newsom vetoes California’s contentious AI “safety” bill SB-1047

Twitter lit up with joyful streams of relief and praise for this decision. Everyone from politicians, economists, researchers, academic luminaries, open source collectives, founders and venture capitalists.

It was a bad bill that lacked the necessary clarity and focus to even begin the task of regulating the nascent field of artificial intelligence.

We can and will do better in finding regulatory frameworks for safety and competitiveness but this bill wasn’t it. It was especially concerning as they say so goes California so goes the world.

I have been banging on about the #FreedomToCompute and math’s crucial role in our constitutional right to free speech in America. This must be considered in all future attempts at regulation in America.

Math and computing power are as essential as speech. In today’s world, they ARE speech. We may speak in natural language, but the way we extend ourselves, build things, and grow as a species is through our tools. Computation is a tool.

These tools are extensions of the human mind. Consider that the first computers were just regular humans counting. We may have started with our fingers and toes as our first tools. And it wasn’t quick progress as the evolution from the abacus to modern computing took us nearly 4,000 years.

We’ve made an astonishing amount of progress in the last hundred years. We’ve gone from thousands of computations per second in the 1940s to 200 quadrillion calculations per second with modern super computers.

Consumer devices are better too. The computer I’m using to write this post has more power than the computers we used to send man to the moon. It’s 100,000 times faster with seven million times more memory.

Alas, as tools get more powerful the powerful get nervous. This isn’t the first bad artificial intelligence bill we’ve seen. We have Europe to thank for that. And it likely won’t be the last.

But defeating SB-1047 is a rare moment of bipartisan cooperation not only in California but across the world as the entire compute space came together to make its voice heard. And Gavin Newsom listened.

We should celebrate this rare consensus as we look towards better policy in our future.

Categories
Internet Culture Startups

Day 1245 and AI: Tool Versus Master

A bad workman blames his tools

Idiom

As we rapidly accelerate the power of our computing tools, machine learning has blossomed into the most heated topic in government policy, business strategy, and popular culture, as artificial intelligence begins to affect everyday life.

The focus on harms, and in particular singularity doomerism, has (ironically) pulled focus from the implications of the seminal “attention is all you need” paper.

Astonishing as it may seem at times, intelligence does not top out at “median human”, but can, and possibly will, go much further.

What a triumph this represents. We will all have access to tools that will enable our entire species to build. More intelligence applied to more problems means more solutions to very real human problems.

I am in Austin Texas for Consensus 2024. I’ll be participating a town hall to discuss how crypto’s mindset of open-source, decentralized computation might help us grapple with who builds, maintains, & owns AI tools.

Policymakers and Silicon Valley executives are both calling for regulation of artificial intelligence as fears grow over its potential harms. But others warn of entrenching a dominant, opaque, centralized Big Tech model and instead advocate for open-source code, decentralized computation and distributed data sourcing. Whom should policymakers listen to? What, if anything, can governments do to help this vital technology evolve in a pro-human way?

Thankfully, we aren’t starting from scratch in building a regulatory framework for artificial intelligence. Code has been treated as speech in Bernstein v. United States Department of State. It would seem like a reasonable precedent to consider algorithms speech as well.

And let us be clear, math and computing power are as essential as speech. In today’s world, they ARE speech. Humans may speak in natural language, but the way we extend ourselves, build things, and grow as a species is through our tools. Computation is a tool. To presume that these tools do harm is to make us bad workmen.

Now of course incumbent powers may try to keep the disruptive and democratizing power of these tools out of the hands of the populace “for their own safety”, but imagine if the first amendment had been frozen in time at the printing press and didn’t protect the internet? We cannot accept permanently lowered standards of fundamental rights.

The 90s era fight for strong encryption enabled a flourishing of digital businesses from finance to e-commerce. We must insist that the freedom to innovate remains the default for U.S. digital policy.

Ultimately, I agree with R Street’s Adam Thierer. He says “fear based narratives that prompt calls for preemptive regulation of computational processes and treat AI innovations ‘as guilty until proven innocent’ are no way to make good policy.”

America does not have a Federal Computer Commission for computing or the internet but instead relies on the wide variety of laws, regulations, and agencies that existed long before digital technologies came along.

R Street Comments on NTIA

If we are to regulate sensibly, let us treat artificial intelligence as we would any other tool and let us do so within the existing framework of our constitutional rights and their interpretations within past precedents.

Categories
Culture

Day 1234 and Intelligence

I’m reasonably intelligent as far as humans go. I’m probably in the top quartile or so of reasoning, processing & other measures of cognition. Not being insecure about my intelligence, I feel perfectly comfortable admitting that I’m an idiot. I’m only human.

Humans just aren’t a terribly bright species. But we are a curious one. We’ve built tools that extend our capabilities significantly. And each new upgrade in our tools helps us achieve more with our meager intelligence. L

We can quibble over whether intelligence is different than achievement but analytical, creative and practical capabilities are things you want to cultivate. We want to cultivate in ourselves and ideally we will want to cultivate in the things we build.

Two men are on a bus on a mountainside side. An sad anxious looking man is staring at the rock face on dark side of the bus side with a thought bubble “the AI took my job” while a smiling happy man on the bright side overlooks scenic views with a thought bubble “the AI took my job.”

Sure we as a species have fought these advances but eventually the benefits of developing ways to pass on and improve intelligence outweighs other fears. Material progress is good.

If you afraid of intelligence greater than your own I realize I have no way to talk you out of that fear. I can argue impartially about the benefits that intelligence has brought us in the past but humans are feeling animals not reasoning animals. The best I can hope for is to coax you to consider the bright side of the bus. Imagine feeling awesome about an AI taking your job. Go ahead and see if your curiosity can consider it.

Categories
Community Internet Culture Politics

Day 973 and Reinforcement

I’ve spent a lot of time this summer thinking about who gets to decide the boundaries of society.

Automation of civic and cultural life has been happening at the speed of capitalism. It’s about to happen at the speed of artificial intelligence’s processing power.

At least during most of techno-capitalism, corporations and governments were still run by humans. You could blame an executive or elected official. What happens when more decisions are snapped into reality by a numerical boundary?

High frequency traders have found many an arbitrage they whittled into nothingness. Who will get whittled away when the machines decide how society should run?

We got a taste of the horrors of treating people like statistics instead of humans during the first Biden era crime bill with mandatory minimum sentencing. And here we are rushing to find new ways to nudge consensus back to hard lines and institutionalization.

I don’t know how we handle virtue in a world without grace. Alasdair MacIntyre’s After Virtue seems prescient. Forgiveness in the face of horrific reality has been the hallmark of humanity’s most enduring religions. But then again so has swift punishment and inscrutable cruelty. Humans are quite a species.

I am, like many others, concerned about reinforcement learning in machine learning and artificial intelligence. How and where we set the boundaries of the machines that determine the course of daily life has been a pressing question since the invention of the spreadsheet.

Marx certainly went on about alienation from our contributions to work. But division of labor keeps dividing. And algorithms seem to only increase the pace of the process.

Categories
Biohacking Medical Startups

Day 971 and Patients Rights With Artificial Intelligence

If you are working in artificial intelligence or medicine I’d like to pleased my case to you. Id just like to pass along a note.

The current “responsible” safety stance is that we should not have AI agents dispense healthcare advice as if they had the knowledge of a doctor. I think this is safetyism and rob’s sick people of their own agency

I have very complicated healthcare needs and have experienced the range of how human doctors fail. The failure case is almost always in the presumption that you will fall within a median result.

Now for most people this is obviously true. They are more likely to be the average case. And we should all be concerned that people without basic numerate skills may misinterpret a risk. Whether it’s our collective responsibility to set limits to project regular people is not a solved problem.

But for the complex informed patient knows they are not average? The real outliers. Giving them access to more granular data let’s them accelerate their own care.

It’s a persistent issue of paternalism in medicine to assume the doctor knows best and the presumption that the patient is either stupid, lying, or hysterical is the norm. It’s also somewhat gendered in my experience.

I now regularly work with my doctors using an LLM precisely so we can avoid these failure cases where I am treated as an average statistic in a guessing game. I’m a patient not a customer after all. I decided my best interest.

A strict regulatory framework constricts access without solving any of the wider issues of access to care for those outside of norms. Artificial intelligence has the capacity to save lives and improve quality of life for countless difficult patients. It’s a social good and probably a financial one too.

Categories
Culture Medical Politics

Day 948 and Assigning Value

What does assigning value mean to you? How do you begin to investigate what is valuable? If someone asked you to value “object X” do you know what tools you would use first to make a measurement?

If I tell you determining value is a cultural problem, you may investigate the problem of value through religious or philosophical frameworks. If I tell you value is an artistic problem, you may use taste in finding value.

If I tell you that assigning value is primarily a computing problem, you may search for weightings, databases and referents to determine value.

So what happens when determining value has to account for multiple or even contradictory frameworks? Which framework assigns the ultimate value? And how do we align them?

Congratulations, you’ve known become an artificial intelligence alignment researcher. I bet you thought that required a doctorate but it doesn’t.

It’s not an entirely intractable problem. The Industrial Revolution found ways to align competing frameworks. We assigned labor value and made currencies to facilitate the exchange of different goods.

Markets can, and do, spring up for all kinds of previously impossible to value things. Capitalism done its best to make cultural value fungible and legible to an agreed upon value. Sure, artisans and artists complain we conclude incorrect values regularly. But we don’t always agree on value.

Generally we’ve found that what can pay for itself survives and what can profit for others thrives.

Not all people are motivated by profit, but we all are motivated to survive. And so we contribute what we believe has value to each other and hope the frameworks of value that others have will align with ours. The balance between the two has held together humanity for sometime.

But deciding on value isn’t the same thing as a thing driving a profit and we have to remember that truth. Between the gaps in the models of what we value is the epsilon of what cannot be calculated.

If you’d like to read a horror story on how assigning fungible value in a database can end up assigning a value to something we humans generally don’t consider interchangeable at all, then I’d go read this piece on how public hospice care’s incentives have been perverted by private equity profit motive.

I don’t always agree with the author of the piece Cory Doctorow. But I think he’s raising a powerful point on how we are assigning value when we overlay competing frameworks.

This is the true “AI Safety” risk. It’s not that a chatbot will become sentient and take over the world – it’s that the original artificial lifeform, the limited liability company, will use “AI” to accelerate its murderous shell-game until we can’t spot the trick

If you aren’t familiar with Doctorow, he’s a powerful voice in right to repair circles, a classical hacker opposed to corporate oligopoly, and a bit of a anarcho-syndicaticalist in his preferred solutions.

I like markets more than governments for most things. More of us can contribute to markets than we can contribute to specialist bureaucracies

But we have assigned value to end of life care inside the convoluted system of profit motives and medical ethics and it’s not the value most of us share on life.

And that’s going to happen a lot more as we get further and further abstracted away from the existing models of value that govern our lives. So remain skeptical when someone tells you that they know what you value. How they assign value might be different than you.