Former (failed) candidate here, state legislative race in a smaller state that's not generally described as competitive. Money = time. When large donors can deploy large amounts of cash, it relieves you of the need to spend tons of time raising money and you can instead spend that time on other things of greater ROI or that can scale your outreach to a larger audience.
There are some efforts to do exactly that, though overall I'd observe that the general level of technological sophistication / early adopter-ism is pretty low in the politician cohort. There are tools for developing campaign plans, direct mail messaging, text campaign content, etc.; we used AI social media ad generation tools to help fill our content pipeline. We developed some internal tools to assist with direct mail fundraising letters to PACs / labor unions / etc.
Not exactly what I was joking about, but interesting nonetheless.
I suppose some amount of adoption by politicians and staffers is good, so that they can see what this new thing is, and crucially, what it is not. But of course it comes at the cost of making a new class of errors. Hopefully these will be contained enough to mainly just serve as a learning experience for them.
As the parent of a teen with ADHD, I find myself comparing my growing-up experiences and anxieties with his. I'm confident that if I were coming of age today, I would probably have been diagnosed with something because I was firmly a couple standard deviations away from the societal mean.
Because we didn't have as extensive diagnoses or therapies back in the 80s compared to now, I had my own phase of wondering what was wrong with me. There weren't any peer or adult role models available to me that really related to my experiences. As a result, there were some difficult years in there...but also, I had to find my own resiliency and ways of mapping my worldview to other people.
Fast forward 40 years. I am conflicted about which is better: to be left to figure it out on your own, or to have a support system that is (at times) overly biased towards leaning on the diagnosis as the explanation. But I can say with high confidence that at least for the coming-of-age years of my child, I am far more thankful that his experiences are different than mine.
"Being a human" is grossly inadequate as a lowest common denominator definition of the needs and experiences of children. Even as broadly discussed as it is, it's still only ~11% of US children and that's still a challenging hill to climb if their peer culture doesn't provide some sort of explanation or incentives for understanding each other.
11% is cited as the percentage that have been diagnosed, so ostensibly the 'true' percentage is even higher. i feel that if such a huge proportion of people (let's say somewhere between 1 in 5 ~ 1 in 10) have an attribute, we shouldn't be treating it as some exceptional thing that requires special care but rather a normal part of what it means to be a human.
it's crazy that there's so much focus on adhd but comparitively little on dyslexia, which by most estimations has similar prevalence but arguably even more child impact. i'm sure there's many more of disadvantages all across the spectrum that we haven't even classified / become aware of.
For me, as much as it pains me to admit this, the sales and account relationship process is just as important of a factor now. I'm at a level where I'm not the end user of most of the infrastructure I purchase for the business, but I'm the one that has to deal with most of the vendor interaction.
Datadog is a pain in the ass. I've got two emails and a voicemail from them just this week. We are not an active customer.
Heroku/Salesforce is also a pain in the ass. It causes enough friction with legal that I'll spend whatever effort it takes to replatform our workload just to not have to have those unending inbound calls.
NS1 was easy-peasy, but post-IBM I now receive a PDF invoice for $50 once per month with no credit card-based billing options and have to remind finance to cut a paper check. I'll be rehoming our DNS as soon as we decide on where to move it to.
tl;dr: the business experience is part of the product
This legal complaint alleges that defendants operating a non-profit entity for the benefit of humanity have committed massive fraud on donors, beneficiaries, and the public. The complaint raises concerns about OpenAI's operation, including its dual structure as a non-profit and a for-profit entity, potential insider dealings, and the exclusion of the general public from its benefits. It claims that OpenAI has used deceptive advertising, unfair competition, and fraud to develop its valuable resource for personal gain.
The complaint highlights OpenAI's mission of benefiting humanity and points out that a narrow group of stakeholders have received commercially invaluable early access to its technology. It also argues that OpenAI's for-profit operations might infringe on copyright and fair use laws, as the technology is built on large datasets, much of which is copyrighted. It accuses OpenAI of breaching trust and fiduciary duties, disrupting legal frameworks, and potentially engaging in willful and wanton negligence by increasing existential risks related to AI.
Finally, the complaint alleges that OpenAI might have engaged in banned political activities, specifically suggesting that the technology may have been used to influence the 2020 US presidential election in favor of the Democratic party.
Wow, that's the most useful ChatGPT summary I've ever come across.
I've never found it particularly useful for most articles which are easy enough to read/skim (the first and last 2 paragraphs will usually tell you what you need), but long complicated legal documents are a whole other matter. This is great.
Works really well for legislation, too. GPT4 can pick up on deltas between bills using markup like <strike> if kept in the document, summarizing changes to the bill as it moves through the legislative process.
The only challenge is chunking the larger bills and synthesizing the larger summary without losing out on possible nuances. Something like California's SB423, for example, is over twice the 8K token limit and that's not even a large bill.
Unfortunately, things like the US Code or Code of Federal Regulations are in the range of 100s of millions of tokens.
ChatGPT picks up 80% of the meaning and rewrites it in beautiful prose. Or maybe another language, in the style of Shakespeare.
On the other hand, if you're in a field where there's an adversarial use of text and the uncomprehended 20% might be used to nullify, contradict or make loopholes in the main body, then relying on ChatGPT is similar to using Tesla Full Self-Driving in a construction zone, near firetrucks, during a snowstorm.
Every summmarization is a choice of salience: what to include and what to leave ou, and how to express something in a different way.
The failure foolishly and misleadingly called “hallucination” is only one manifestation of an attribution error. If your summarizer leaves out something very important because it doesn’t understand it the result will be quite misleading.
For your average web text which these days is 90% filler and not important anyway, this is no big deal. This particular lawsuit appears the same. But for anything important, I wouldn’t trust it.
In my experience it’s generally accurate when summarizing content provided in the prompt context. Where it can run into trouble is “recalling” (if you can call it that) content that it was trained on.
You just copy/paste the text of the complaint (this is why you hear many complaining about maximum context size... for some use cases you want to feed in a lot of text)
A friend of mine works at a law firm that exclusively handles train injury cases. It happens a LOT more than you'd expect, primarily at crossings. Hearing-handicapped or people blasting their speakers who don't hear the horn; faulty crossing gates; distracted drivers; and so on.
I don't personally know the OP, but his profile isn't that of a passing crypto fanboy hustling to attract a followers list. I'd wager that he probably doesn't care much if some pro-Elon HN folks don't agree with his views on capitalism or proper vs. improper ways to migrate a follower base between platforms.
He's contributed to the software industry longer than most people here. That's not to say that anyone is owed anything in deference or respect, but "he's seen some shit." It's hypocritical to applaud what you believe to be pro-free-speech changes at Twitter while criticizing somebody for the manner in which somebody communicates their decision to deprioritize that platform.
If somebody has 50k followers, I imagine they would have reasons to communicate to that audience why they're not going to be as visible on their current platform and how to go about finding their content in the future.
Seems relevant to me. The article is from your organization, and that information is something I wouldn't have known if not for the prior commenter's comment.
Ten years or so ago, I was participating in a small business roundtable discussion with one of our state senators. At the time, I ran a consumer research agency and would often have multinational projects involving consumer data collection in both the US and EU; this is before GDPR had become ratified, but Safe Harbor was failing and there was ambiguity about what the future state would look like.
Of the 15 or 20 business owners in the room, I was the only "pro privacy" voice. People were very focused on what would be the perceived additional cost of complying with any GDPR-style rules in the US, and weren't yet thinking about the negative effects of having different privacy rules in different markets. "Different markets have different rules all the time," in short.
I maintain that it would be less complicated, less expensive, and more human-friendly to use data privacy rules as globally universal as can be achieved. There will always be capitalism leeches that drain money through arbitrage between the policy gaps, yes, but it would help.
(Also: there is zero chance this gets through the current US Senate. Would never clear filibuster.)
I'm mostly just projecting based on the current 48+2+50 state of the Senate where virtually everything gets held up. If the Democrats brought it forward, I would expect the Republicans to filibuster just on principle.
Slight historical aside here: for those too young to have been politically aware before the Obama era, the filibuster (a) is in the modern form a relatively recent invention (b) was rarely used prior to the Mitch McConnell era. McConnell and the Fox News generation of the Republican Party turned what was meant to be a tool for a last-resort veto into a sledgehammer continuously used to bludgeon the other side.
> I maintain that it would be less complicated, less expensive, and more human-friendly to use data privacy rules as globally universal as can be achieved.
I think this is a bit naive. As someone who has had to dwell a lot on the specific nuances of German privacy laws vs GDPR or South Korea's, I have come to the conclusion that conflicting privacy laws are a designed feature.
I think lawmakers certainly have consumer protection as one of their goals, most privacy legislation has many features intended to benefit domestic industries at the expense of foreign ones. Or to benefit national security in some way (such as requirements for certain types of data to be stored on servers inside the country).
Even if the US was to homogenize with GDPR in some way, I wouldn't doubt that the EU would fast follow with a slightly different spin on it just to give US tech companies an extra set of hoops to jump through.
In a way, this is already how safety regulations work in the automobile industry.
I agree that we're not going to see a US privacy framework that's identical to GDPR and where all players have the same obligations and enforcement mechanisms. What is extremely problematic, IMHO, is the US having _no_ privacy framework to speak of while the rest of the world does. Beyond HIPPA and COPPA (and CCPA if you happen to live in Cali), there's really not much recourse for US citizens besides their collection of company-paid credit monitoring after each security breach.
If one outcome of GDPR is that 10-15 years later, the US adopts some sort of national privacy framework that motivates industry to reevaluate their data monetization business models, that's a good outcome.
