> Judge Yvonne Gonzalez Rogers ordered the iPhone-maker to allow developers to steer users to alternative methods of paying for services or subscriptions offered in the App Store. The company also can no longer impose fees in such scenarios or restrict the ability of software-makers to offer links or otherwise communicate alternate payment options with consumers.
The ruling means this starts immediately it seems, as I cannot see a date listed anywhere.
So : make a free app. Ask people to buy the non-free version on a website. Apple gets 0. No revenue at all for providing stable apis and sdks? How’s that sustainable?
The app store itself would be a gigantic loss leader that has to be paid for by iphone sales.
If console makers have to do the same, consoles triple in price
Do you also believe that Apple and Microsoft are entitled to a 30% cut when you subscribe to Spotify, Netflix or buy a game on Steam on a desktop device? Right now Apple and Microsoft both get 0, no revenue at all for providing stable apis and sdks, how's that sustainable?
Where does it stop, should Microsoft, Nvidia, and Intel get a cut when I place an order on my PC that uses APIs and SDKs provided by those companies? Does this entitlement extend to anyone who facilitated the transaction, like my power company and ISP?
...Apple makes revenue on the devices? On any monthly subscriptions you take with them? On the developers paying for licenses?
And let's not act as if people would use iOS if they couldn't have their apps. It is so much of a net gain to Apple to allow people to develop their apps there's a good chance it's indirectly profitable without any direct revenue from IAP's.
It's actually doing really well but it has not much to do with them being a red state and everything to do with them having an aging or tourist transit population that needs to get between cruise lines and theme parks.
Besides those cases there are not a lot of reasons to be transiting that route, or at least not nearly compared to the number of folks that would want to transit SF <-> LA. Totally different needs that are easier to address, and less pressure to prove out value / fewer digestible alternatives.
The problem with local infrastructure projects is their responsibility and execution fall to local community leaders (in this case local = state of California). These leaders are far less efficient and impactful than the federal government, a bar which is already way too low. Nearly every city in the US has had recent collisions with corruption, blatant mismanagement, ideological forces co-opting process, etc. and so most local infrastructure projects are doomed from the start.
The problem with high speed rail, then, is that it is always going to be a local (=state) infrastructure project. America is far, far, far too large for high speed rail to be feasible at a national level, and so we have invested in airports. This is largely a success; much of the year you can travel from NYC to Miami (~1200 mi, roughly UK -> Spain) for $100-200 in a few hours. There are, of course, many issues with air travel that we are still working on, but unless there is a breakthrough to make supersonic land travel affordable, we are stuck with air travel at a national level.
But where do we go from here? We know the federal process is too bloated to succeed with infrastructure projects, and when it is forced to it ends up being prohibitively expensive. We know the state process is doomed to fail and similarly be very expensive. We have already tried privatizing it and failed, and even when subsidizing private industry we get subpar results at best. What options are even left at this point?
We could actually just pay more money to public governmental employees, allow them to build careers (and raise families), and stop contracting out everything. This would also increase competition for such roles.
How we got here is ultimately, especially for CA, land and legal cost, cost of living, and public employee salary relative to private salaries. That’s endemic to major metro areas in the US, with some small exceptions, but especially true in CA.
Before it was shut down, we did see a real reversal with 18F in getting things through, for software projects. Of course, they weren’t even being paid industry wages there.
> We could actually just pay more money to public governmental employees
When their pensions are taken into account, government employees are ridiculously overcompensated already. They’re basically minor nobility at this point.
I have done the numbers for myself, and even taking pension and treating an extra week of vacation as a 1/52 pay raise, doing a software job for the government would be a huge step down for me financially. Basically gives me 0% chance of ever buying a home where I currently live
> me 0% chance of ever buying a home where I currently live
The housing crisis is the Everything Crisis. It’s destroying competence of government services, as no one ambitious will accept this fate.
It destroyed competitiveness of our industry because you can’t pay a worker less than it costs to rent.
It’s causing apathy and rise in extremist views among younger population as they realise they have no path to dignified future.
I am hoping that China does well for itself and one day we can just consider them as an example of competent and coherent governance and sort out our shit.
I guess we'll see how I'm doing in 20 years to compare, but as of now, my dad is a former California public employee on a pension who had to move to the middle of the desert to stay within his budget, does all domestic work himself, and regularly needs to ask me for money. "Minor nobility" is a laughable characterization of his status in life.
For what it's worth, it's also quite illuminating these days to compare injury history with him. We've had a lot of the same stuff happen, but when it happens to me, he always overestimates the effect because he's still suffering from never recovering properly thanks to a lifetime of no-premium but shitty Kaiser HMO treatment that used the lowest-cost option for literally everything, whereas I had to pay for stuff but at least got the latest available surgical procedures and proper supervised rehab, so my injuries actually healed and I'm not just suffering for the rest of my life like he does.
Then you would have two retirees for every three working age people. And given that only ~70% of working age Americans are actually working, it would mean three non-working adults for every two adults with jobs. Either the retirees would have a really low standard of living, or workers would get much smaller share of the value they create than they currently do.
If you are not already retired or close to retirement, you should assume that the normal retirement age is ~70 years. Anything lower cannot be sustained with the current demographic structure.
You're right and everyone -- including government contractors -- has been quietly shouting this for ages, but existing efforts have already been ideologically co-opted and new efforts are too brash and can't wrap their head around spending more to save more.
Also 18F was not truly shut down and still lives on at GSA, albeit only the worst parts. They're hard at work adding as much red tape as they can imagine to procurement of software and recently boasted about their collective organization answering a whopping 1200 emails in a month.
I suspect we're dealing with the fallout of the loss of an American nomos (shared values, traditions, and moral principles formalized into law, custom, and convention) -- the very issue John Adams wrote about in a 1798 letter:
While our country remains untainted with the principles and manners which are now producing desolation in so many parts of the world; while she continues sincere, and incapable of insidious and impious policy, we shall have the strongest reason to rejoice in the local destination assigned us by Providence. But should the people of America once become capable of that deep simulation towards one another, and towards foreign nations, which assumes the language of justice and moderation, while it is practising iniquity and extravagance, and displays in the most captivating manner the charming pictures of candour, frankness, and sincerity, while it is working diligently under the cover of these pleasing appearances and employing the most insidious and base artifices, this country will be the most miserable habitation in the world. Because we have no government, armed with power, capable of contending with human passions, unbridled by morality and religion. Avarice, ambition, revenge and licentiousness would break the strongest cords of our Constitution, as a whale goes through a net. Our Constitution was made only for a moral and religious people. It is wholly inadequate to the government of any other.
When moral virtue is not valued, who will rise to the top and be elected to public office but non-virtuous people?
Failing the development of something like what John Adams is referring to, I fear that the only way "forward" (if it can be called that) is a different form of government, in which individual liberties will be greatly reduced or denied altogether.
My personal opinion -- it's a spiritual problem that needs a spiritual solution. Pray for the nation.
Or we can just keep trying to run through molasses :)
>But should the people of America once become capable of that deep simulation towards one another, and towards foreign nations, which assumes the language of justice and moderation, while it is practising iniquity and extravagance, and displays in the most captivating manner the charming pictures of candour, frankness, and sincerity, while it is working diligently under the cover of these pleasing appearances and employing the most insidious and base artifices
It's always amazing to me that slave owning men on stolen land could not see the hypocrisy of such a statement
John Adams in particular was not a slave owner, and he called slavery an "evil of colossal magnitude." So he and those like him (and there were many) shouldn't be condemned in any sense along that line.
I agree that slavery and ill treatment of Native Americans were egregious problems. But on both issues, there were prominent voices speaking out in favor of what was right, including among the founding fathers -- Ben Franklin, Alexander Hamilton, John Jay, etc. Also, William Penn (founder of Pennsylvania) was an awesome example of someone who insisted on honest, fair dealings with Native Americans, and he won their respect in a big way.
It is hypocrisy to preach about freedom and the equality of man while also owning slaves and believing some men to be unequal. It's not that this was unknown at the time either. Abolitionism was not a new idea in 1798.
Well sure, Adams in particular maybe not, but for him to warn about this hypothetical scenario while it was ostensibly unfolding in front of his eyes is still something i can find a bit ridiculous.
> for him to warn about this hypothetical scenario while it was ostensibly unfolding in front of his eyes
Well why wouldn't he? See something, say something.
I'm sure he saw direct reasons to be worried about people becoming ungovernably immoral, and I bet slavery was one of those reasons -- after all, he did believe it to be a colossal evil.
That's a misquote - that phrase doesn't appear in the letter. I don't think that idea was present either. Could you clarify which phrase(s) conveyed that to you?
If anything, I think he felt compelled to write what he did (not only in this letter) because he felt a sense of imminent danger -- that if the people individually and collectively failed to rise to the high calling of good moral character, the new republic would not last. Remember, the longevity of the United States was by no means a foregone conclusion at the time -- the US Constitution had been ratified a mere 10 years before he wrote this letter, and Adams himself had just been elected as the second president the year before.
(All of which doesn't make his thoughts inapplicable to our time.)
So I think the best argument against this is that the US has done massive infrastructure projects in the past. Some cities had (and some still have) decent public transportation and we've done large canals and even the interstate highway project. I've also seen local infrastructure (and I mean actually local, on the city/county level) be very impactful in the community.
My experience is that unless the community really wants the infrastructure it won't get done and I've seen a lot of opposition to forms of transport that aren't just roads (with the most extreme case of a vocal group in a town not wanting a rail line into the city because they were worried people from the city would come out to their town. They literally didn't want too easy of transportation into/out of the city).
Your theory does not explain why the state of California is capable of discharging $25 billion into freeway construction annually. It obviously cannot be the case that US states are incapable of executing public works projects.
It is so, so, so much easier to build freeways, roads, and highways than rail systems.
It is also absolutely necessary to have roads to move people and goods as not everyone and everything can be transported on rail or even is near rail (even post-HSR).
This is not really the point. Again, your hypothesis lacks explanatory powers. Virtually all of the works undertaken so far on the "rail" project have been roads: bridges, trenches, and relocations of car stuff. So if California can spend tens of billions every year on car stuff through the caltrans budget, it should be more than capable of moving a few of those dollars over the get the HSR project finished. It's a matter of political leadership.
> It is also an income center. California _makes_ money from spending on roads because of job creation (=tax revenue), taxes on vehicle sales, taxes on fuel, tolls, freight fees, and more: https://www.calbike.org/there-is-no-deficit-in-californias-t...
Please provide a reference for CA making money by building roads. Road building is foundational to other economic activity, but AFAIK, it's not turning a profit (nor should it be expected to). Unfortunately, we have a double standard where we consider public transport to be a failure if it doesn't generate a profit.
CA has built tens of miles, probably hundreds, maybe even into the thousands of miles of tracked transit in the time HSR has been building. CA can certainly do what it wants - or maybe do what it needs.
That doesn’t make sense. California is bigger and richer than all but the largest European countries. It has some of the top universities in the world. Why shouldn’t it have a higher caliber of public official than say Denmark?
> It has some of the top universities in the world. Why shouldn’t it have a higher caliber of public official than say Denmark?
Many officials in the current administration who are wreaking havoc on the US economy went to the top universities, like Harvard, Columbia, etc. Attending those universities doesn't imply responsibility or sense of duty to public service at all.
CA does have "high caliber" public officials, but it also produces very corrupt ones. As everyone knows, CA has exported many from both categories to the national stage.
CA is not like Denmark in many ways. California has huge wealth inequality, and its economy and governance are heavily influenced by plutocrats from industries that grow via massive scaling properties, like tech, entertainment, and agriculture.
Two of those industries have a history of labor exploitation, and the other has been actively trying to snuff out the power and leverage it mistakenly gave to its workers over the past several decades.
Without a doubt, Denmark has problems also, and it has plutocrats, but its politicians hold them more accountable to the populace that CA does.
Singapore and Denmark are far more similar to each other than either is to California. Not only are they far smaller, but both countries socially invest heavily in their citizens' quality of life, and by American standards both are egalitarian.
Singapore carefully manages housing, with the majority of the population living in public housing. Of course, both countries also have excellent public transportation.
Why are the politicians in California different than the ones in Denmark? And are those reasons related to it being a sub-national entity, which seems to have been the argument being made above.
> Why are the politicians in California different than the ones in Denmark?
> And are those reasons related to it being a sub-national entity, which seems to have been the argument being made above.
I'm sure it plays a role, but there are tons of convolved factors? They're different places, and your question is too non-specific to have a single answer.
Calling the leaders of the State of California, the 4th largest economy in the world, "local" and less efficient than the federal government is interesting.
Additionally the U.S. Government requires that the states have balanced budgets, a requirement that is not stipulated for federal national spending. Unfunded spending obligations are not a problem, the federal government just raises the debt ceiling.
HSR is perfectly viable for the east. decades of steady investment could have built a massive network out through chicago by now, and with a few more we could be closing in on the last gap in the network between kansas city and las vegas.
This would work if we could run state governments like mini "countries" but there is not enough interest from the electorate and thus very little oversight so they end up in a bad way.
I would like very much for California's state legislature to be on par with, say, South Korea (similar population size & dispersion with great bullet train infrastructure). Unfortunately it operates much more like Thailand or El Salvador pre-"throw everyone in jail".
> This would work if we could run state governments like mini "countries" but there is not enough interest from the electorate and thus very little oversight so they end up in a bad way.
Which is circular; there's less attention because we don't.
> I would like very much for California's state legislature to be on par with, say, South Korea (similar population size & dispersion with great bullet train infrastructure).
About 1/4th the land area of California and much more .. "square". The long line in South Korea is 275 miles long; versus roughly 800 miles for the California HSR system. And California's system is facing a public that's much less accustomed to and receptive to rail, and endpoints that basically require you to have a car anyways.
We should have spent all this money on making local rail awesome. It'd make a much bigger difference in day to day life, would pay off quicker, and would prepare the ground for doing HSR.
they have more efficient central planning, less red tape, and they don't let people like Elon Musk come in and derail the project with dubious and irrelevant alternatives
It really all comes down to civil vs. common law. In America, anyone can hold you over a barrel in court, indefinitely, for any or no reason, until you agree to their extortion.
unfortunately it is worse than that.. a basic and non-trivial State of California project was being done in an office near me, and I knew some of the people. So I saw a bit about how it progressed.. lots of requirements, lots of people from multiple unrelated and slightly competitive groups. The work was mostly intellectual assesment and evaluations with a lot of reports. The thing was funded at professional values.
About six months of work with lots of progress meetings with State of California bureaucrats to "keep them informed" .. and lo-and-behold.. as the required deadlines started getting closer, the State reps changed requirements, made amendments to deliverables.. the last three weeks, even MORE change orders "non-negotiable" .. I have never in my life seen major requirements changed on a multi-party project in the last weeks of a deadline like that. It was like the bureaucrats drank a lot of something, felt the "excitement" and HAD to change things to be "involved and hands on" .. it was STUPID and caused DAMAGE. There was no choice -- the State was paying.
That is how they do things in "infinite income" Sacramento ?
What does that have to do with civil vs. common law? There isn't anything inherent in rule of law based on cases and precedent that should restrict infrastructure projects.
I get so tired of hearing that HSR doesn't make sense federally.
The entire country east of the Mississippi has comparable population and density to western Europe. Plus the sheer size and lack of people sure doesn't stop the federal government from maintaining thousands of miles of interstate through vast swaths of nowhere.
If you look at it by map there's not much more interstate highway than railway.
Interstate highway is also far cheaper to build and maintain than you would imagine. It took around $114 billion to build interstate across the entire country, while the high speed rail project in just California is already upwards of $128 billion.
On Europe that is kind of exactly my point, high speed rail in Europe is built and maintained by federal governments with high levels of participation, interest and oversight from their population. This can never happen in the US because our federal government has to oversee a very wide amount of area and states are not so autonomous and self-governed that their populations primarily interest themselves with their state governments.
So are we going to pretend Palantir, Anduril, SpaceX do not exist with their armies of 21 year olds fresh out of college getting fast tracked for SCI and working with secrets far more dangerous than nuclear?
Also by definition everyone who receives an SCI clearance would have 0 experience working with anything of that nature. Nearly everyone who gets an account on JWICS (btw there are 30k+ accounts on JWICS, not exactly a small group) would have no experience working with the information hosted there.
I don't follow the line of reasoning in this article and honestly expected more than tabloid style reporting from NPR. Imagine the converse, if there were swathes of people with experience working with nuclear weapons or more dangerous military technology before getting any type of clearance from their government.
The issue is that, at least from what has been reported, DOGE hasn't gone through standard protocol to get access. No background checks, no education, no vetting and no oversight.
And what exactly is the "need to know" for DOGE here?
You are confusing secret+ classification with CUI.
Above I specifically refer to SCI, the designation for things of dire national security impact, which quite literally mandates that information be fully compartmentalized and unattainable to the public. Above that we have SAPs or special clearances (e.g. Q for Nuclear, YW for White House).
It is not possible to work with nuclear weapons, serious satellites, advanced drone weaponry, or similar without first getting an SCI clearance. In fact all of these are prime examples of their specific marking compartments. There are even regulations below this (e.g. ITAR) that control who gets to work on military technology in general.
there is a doctrine of born secret. and it is a contingency.
if i, or you, or anyone, was to independently realize how to increase yield of a weapon, or create a nuclear isomer weapon, it is secret upon conception.
knowledge is discoverable, and cant be stopped, only access to the tools of production can be stopped.
that is why you CANT stop motivated persons from making independent discovery, you have to prevent your adversary from knowing what you do/dont know, and what you are/are not doing.
This might be a bit off-topic but is this custom-built documentation or did you use a template? I've been on the hunt for pretty documentation tools but have only come up with Mintlify.
Is it custom all the way down or based on something like starlight? I really like Mintlify but $200 per doc site hurts for internal tools.
I can see a lot of merit in custom building though especially with making it easier to dump it into `llms.txt` or exposing a search for AI. Hoping that's where DeepWiki is headed :)
When this was up yesterday I complained that the refusal rate was super high especially on government and military shaped tasks, and that this would only push contractors to use CN-developed open source models for work that could then be compromised.
Today I'm discovering there is a tier of API access with virtually no content moderation available to companies working in that space. I have no idea how to go about requesting that tier of access, but have spoken to 4 different defense contractors in the last day who seem to already be using it.
"Alignment with who?" has always been a problem. An AI is a proxy for a reward function, a reward function is a proxy for what the coder was trying to express, what the coder was trying to express is a proxy for what the PM put on the ticket, what the PM
put on the ticket is a proxy for what the CEO said, what the CEO said is a proxy for shareholder interests, shareholder interests are a proxy for economic growth, economic growth is a proxy for government interests.
("There was an old lady who swallowed a fly, …")
Each of those proxies can have an alignment failure with the adjacent level(s).
And RLHF involves training one AI to learn human preferences, as a proxy for what "good" is, in order to be the reward function that trains the actual LLM (or other model, but I've only heard of RLHF being used to train LLMs)
More accurate to call it “alignment for plebes and not for the masters of the plebes”. Which I think we all kind of expect coming from the leaders of our society. That’s the way human societies have always worked.
I’m sure access to military grade tech is only one small slice in the set of advantages the masters get over the mastered in any human society.
"Protecting the world" would require a common agreement on morals and ethics. OpenAI shitting it's pants when asking how to translate "fuck", which OpenAI refused for a very long time, is not a good start.
Morals and ethics are different and I would not want the US to be "protecting the world" with their ridiculous ethics and morals.
I've always thought that if a corporate lab achieves AGI and it starts spitting out crazy ideas such as "corporations should be taxed," we won't be hearing about AGI for a while longer due to "alignment issues."
Can you explain the difference between taxing the corporation itself vs taxing the executives, board members, investors, and employees directly (something that already happens)?
I really don't know where to begin answering this.
It is generally accepted that business profit is taxed. Meanwhile, there are entire industries and tax havens set up to help corporations and their executives avoid paying taxes.[0]
However, the crux of my comment was not about the vagaries of corporate taxation, it was simply about "AI alignment" being more about the creators, than the entire species.
> I really didn't expect so much paperclip production growth this quarter!
>> How'd you do it?
> I don't know the details. ChatGPT did it for me, this thing's amazing. Our bonuses are gonna be huge this year, I might even be able to afford a lift kit for my truck.
It's "tier 5", I've had an account since the 3.0 days so I can't be positive I'm not grandfathered in, but, my understanding is as long as you have a non-trivial amount of spend for a few months you'll have that access.
(fwiw for anyone curious how to implement it, it's the 'moderation' parameter in the JSON request you'll send, I missed it for a few hours because it wasn't in Dalle-3)
I just took any indication that the parent post meant absolute zero moderation as them being a bit loose with their words and excitable with how they understand things, there were some signs:
1. it's unlikely they completed an API integration quickly enough to have an opinion on military / defense image generation moderation yesterday, so they're almost certainly speaking about ChatGPT. (this is additionally confirmed by image generation requiring tier 5 anyway, which they would have been aware of if they had integrated)
2. The military / defense use cases for image generation are not provided (and the steelman'd version in other comments is nonsensical, i.e. we can quickly validate you can still generate kanban boards or wireframes of ships)
3. The poster passively disclaims being in military / defense themself (grep "in that space")
4. it is hard to envision cases of #2 that do not require universal moderation for OpenAI's sake, i.e. lets say their thought process is along the lines of: defense/military ~= what I think of as CIA ~= black ops ~= image manipulation on social media, thus, the time I said "please edit this photo of the ayatollah to have him eating pig and say I hate allah" means its overmoderated for defense use cases
5. It's unlikely openai wants to be anywhere near PR resulting from #4. Assuming there is a super secret defense tier that allows this, it's at the very least, unlikely that the poster's defense contractor friends were blabbing about about the exclusive completely unmoderated access they had, to the poster, within hours of release. They're pretty serious about that secrecy stuff!
6. It is unlikely the lack of ability to generate images using GPT Image 1 would drive the military to Chinese models (there aren't Chinese LLMs that do this! even if they were, there's plenty of good ol' American diffusion models!)
I'm Tier 4 and I'm able to use this API and set moderation to "low". Tier 4 only requires a 30 day waiting period and $1,000 spent on credits. While I as an individual was a bit horrified to learn I've actually spent that much on OpenAI credits over the life of my account, it's practically nothing for most organizations. Even Tier 5 only requires $5,000.
OP was clearly implying there is some greater ability only granted to extra special organizations like the military.
With all possible respect to OP, I find this all very hard to believe without additional evidence. If nothing else, I don't really see a military application of this API (specifically, not AI in general). I'm sure it would help them create slide decks and such, but you don't need extra special zero moderation for that.
> With all possible respect to OP, I find this all very hard to believe without additional evidence. If nothing else, I don't really see a military application of this API (specifically, not AI in general). I'm sure it would help them create slide decks and such, but you don't need extra special zero moderation for that.
I can't provide additional evidence (it's defense, duh), but the #1 use I've seen is generating images for computer vision training mostly to feed GOFAI algorithms that have already been validated for target acquisition. Image gen algorithms have a pretty good idea of what a T72 tank and different camouflage looks like, and they're much better at generating unique photos combining the two. It's actually a great use of the technology because hallucinations help improve the training data (i.e. the final targetting should be invariant to a T72 tank with a machine gun on the wrong side or with too many turrets, etc.)
That said, due to compartmentalization, I don't know the extent to which image gen is used in defense, just my little sliver of it.
We can talk about it here, they put out SBIRs for satellite imagery labeling and test set evaluation that provide a good amount of detail into how they're using it.
There are plenty of fairly mundane applications for this sort of thing in the military. Every base has a photography and graphic design team that makes posters, signs, PR materials, pamphlets, illustrations for manuals, you name it. Imagine a poster in the break room of a soldier in desert gear drinking from his/her canteen with a tagline of "Stay Alive - Hydrate!" and you're on the right track.
I'm not aware of the moderation parameter here but these contractors have special API keys that unlock unmoderated access for them, they've apparently had it for weeks.
Think of all the trivial ways an image generator could be used in business, and there is likely a similar use-case among the DoD and its contractors (e.g. create a cartoon image of a ship for a naval training aid; make a data dashboard wireframe concept for a decision aid).
Input one image of a known military installation and one civilian building. Prompt to generate a similar _civilian_ building, but resembling that military installation in some way: similar structure, similar colors, similar lighting.
Then include this image in a dataset of another net with marker "civilian". Train that new neural net better so that it does lower false positive rate when asked "is this target military".
You might not believe it but the US military actually places a premium on not committing war crimes. Every service member, or at least every airman in the Air Force (I can't speak for other branches) receives mandatory training on the Kunduz hospital before deployment in an effort to prevent another similar tragedy. If they didn't care, they wouldn't waste thousands of man-hours on it.
> On 7 October 2015, President Barack Obama issued an apology and announced the United States would be making condolence payments of $6,000 to the families of those killed in the airstrike.
Bombs and other kinds of weapon system which are "smarter" have higher markup. It's profitable to sell smarter weapons. Dumb weapons is destroying the whole cities, like Russia did in Ukraine. Smart weapons is striking a tank, a car, an apartment, a bunker, knowing who's there and when — which obviously means less % of civilian casualties.
The very simple use case is generating mock targets. In movies they make it seem like they use mannequin style targets or traditional concentric circles but those are infeasible and unrealistic respectively. There's an entire modeling industry here and being able to replace that with infinitely diverse AI-generated targets is valuable!
I don't really understand the logic here. All the actual signal about what artillery in bushes look like is already in the original training data. Synthetic data cannot conjure empirical evidence into existence, it's as likely to produce false images as real ones. Assuming the military has more privileged access to combat footage than a multi-purpose public chatbot I'd expect synthetic data to degrade the accuracy of a drone.
Generative models can combine different concepts from the training data. For example, the training data might contain a single image of a new missile launcher at a military parade. The model can then generate an image of that missile launcher hiding in a bush, because it has internalized the general concept of things hiding in bushes, so it can apply it to new objects it has never seen hiding in bushes.
I'm not arguing this is the purpose here but data augmentation has been done for ages. It just kind of sucks a lot of the time.
You take your images and crop, shift, etc them so that your model doesn't learn "all x are in the middle of the image". For text you might auto replace days of the week with others, there's a lot of work there.
Broadly the intent is to keep the key information and generate realistic but irrelevant noise so that you train a model that correctly ignores the noise.
You don't want to train your model identifying some class of ship to base it on how choppy the water is, just because that was the simple signal that correlated well. There was a case of radiology results that detected cancer well but actually was detecting rulers in the image because in images with tumors there was often a ruler so the tumor could be sized. (I think it was cancer, broad point applies if it was something else).
If you're building a system to detect something, usually you need enough variations. You add noise to the images, etc.
With this, you could create a dataset that will by definition have that. You should still corroborate the data, but it's a step ahead without having to take 1000 photos and adding enough noise and variations to get to 30k.
I can get an AI to generate an image of a bear wearing a sombrero. There are no images of this in its training data, but there are bears, and there are images of sombreros, and other things wearing sombreros. It can combine the distributions in a plausible way.
If I am trying to train a small model to fit into the optical sensor of a warhead to target bears wearing sombreros, this synthetic training set would be very useful.
Same thing with artillery in bushes. Or artillery in different lighting conditions. This stuff is useful to saturate the input space with synthetic examples.
Unreal, Houdini and a bunch of assets do this just fine and provide actually usable depth / infrared / weather / fog / TOD / and other relevant data for training - likely cheaper than using their API
See bifrost.ai and their fun videos of training naval drones to avoid whales in an ethical manners
well considering an element of their access is the lifting of safety guardrails, I'd assume the scope includes, to some degree, the processing or generation of nsfw/questionable content
Interesting. Let's say we have those and also 30k real unique images, my guess is that real ones would have more useful information in them, but is this measurable? And how much more?
The model they're training to perform detection/identification out in the field would presumably need to be much smaller and run locally without needing to rely on network connectivity. It makes sense, so long as the openai model produces a training/validation set that's comparable to one that their development team would otherwise need to curate by hand.
Vastly oversimplified but for every civilian job there's an equivalent military job. Superficially, the military is basically a country-sized self-contained corporation. Anywhere that Wal-Mart's corporate office could use AI so could the military.
That's very outdated, they're absolutely supposed to be at the Empire State Building with baseball caps now. See: ICE arrests and Trump's comment on needing more El Salvadoran prison space for "the homegrowns"
Show me a tunnel underneath a building in the desert filled with small arms weapons with a poster on the wall with a map of the United States and a label written with sharpie saying “Bad guys here”. Also add various Arabic lettering on the weapons.
All I can think of is image generation of potential targets like ships, airplane, airfield and feed them to their satellite or drones for image detection and tweak their weapons for enhance precision.
I think the usual computer vision wisdom is that this (training object detection on generated imagery) doesn't work very well. But maybe the corps have some techniques that aren't in the public literature yet.
My understanding is the opposite, see papers for "synthetic" data training. They use a small bit if real data to generate lots of synthetic data and get usable results.
The bias leans towards overfitting the data, which in some use cases - such as missile or drone design which doesn't need broad comparisons like 747s or artillery to complete it's training.
Kind of like neural net back propogation but in terms of model /weights
In 2024, the Pentagon carved out an exception for themselves on the Huawei equipment ban [0]
I would imagine defense contractors can cut deals for similar preferential treatment with OAI and the like to be exempt from potentially copyright-infringing uses of their API.
Just ask Microsoft about Tay. On the one hand, I understand why you want some censoring in your model, on the other, I think it also cripples your models in unexpected ways, I wonder if anyone's done such research, compare two models by the same source training data, one with censoring of offensive things, the other without. Which one provides more accurate answers?
"GPT-4o is now available as part of Azure OpenAI Service for Azure Government and included as part of this latest FedRAMP High and DoD IL4/IL5 Authorization."
...we have everything setup in Azure but are weary to start using with CUI. Our DoD contacts think it's good to go, but nobody wants to go on record as giving the go-ahead.
Ah by “it” I meant OpenAI commercial. Azure OpenAI can handle CUI Basic.
They also have a deployment on SIPR rated for secret.
Anything higher, you need a special key but AWS Bedrock has Claude up on C2S.
That being said both Azure OpenAI and AWS Bedrock suck for many reasons and they will by default extend your system boundary (meaning you need to extend your ATO). Also, for CUI, it has the P-ATO from JAB, not many agency specific ATOs, which means you will probably need to submit it thru your agency sponsor.
Gotcha. We happen to be on government Azure as a contractor, which took years to secure (and one reason our execs want to be beyond sure everything is locked down)
Have they given a reason for being hesitant? The whole point of IL4+ is that they handle CUI (and higher). The whole point of services provided for these levels is that they meet the requirements.
This is on purpose so OpenAI can then litigate against them. This API isn't about a new feature, it's about control. OpenAI is the biggest bully in the space of generative AI and their disinformation and intimidation tactics are working.
Honestly in my anecdotal experience so far, it's been about the same as smaller and cheaper tools like HiDream that we can host and control ourselves.
I'm also getting a surprisingly high number of refusals for everything from stylistic representation to topics that are only lightly controversial like government/military.
e.g. I'm unable to get this to generate a drone "target" for infantry practice. This is surprising given OpenAI's ongoing IL4 efforts. I don't know why they would take this sort of Anthropic-esque "ethics" stance, especially given there are already several companies selling the above to the DoD.
It's a huge failure mode if people end up using CN-developed open source AI models for US government & military use cases because software engineers at our own AI labs are taking holier-than-thou ethics stances.
I think the most important thing to note here, perhaps more so than the context window, is that this exposes some serious flaws in benchmarks. Per benchmarks, Maverick is competitive only with older models like GPT-4o or Gemini 2.0 Flash, and not with anything in the last few months (incl. reasoning models).
The benchmarks are awful. No disrespect to the people who worked to make them, nothing is easy. But I suggest going through them sometime. For example, I'm currently combing through the MMMU, MMMU-Pro, and MMStar datasets to build a better multimodal benchmark, and so far only about 70% of the questions have passed the sniff test. The other 30% make no sense, lead the question, or are too ambiguous. Of the 70%, I have to make minor edits to about a third of them.
Another example of how the benchmarks fail (specifically for vision, since I have less experience with the pure-text benchmarks): Almost all of the questions fall into either having the VLM read a chart/diagram/table and answer some question about it, or identify some basic property of an image. The former just tests the vision component's ability to do OCR, and then the LLM's intelligence. The latter are things like "Is this an oil painting or digital art?" and "Is the sheep in front of or behind the car" when the image is a clean shot of a sheep and a car. Absolutely nothing that tests a more deep and thorough understanding of the content of the images, nuances, or require the VLM to think intelligently about the visual content.
Also, due to the nature of benchmarks, it can be quite difficult to test how the models perform "in the wild." You can't really have free-form answers on benchmarks, so they tend to be highly constrained opting for either multiple choice quizzes or using various hacks to test if the LLM's answer lines up with ground truth. Multiple choice is significantly easier in general, raising the base pass rate. Also the distractors tend to be quite poorly chosen. Rather than representing traps or common mistakes, they are mostly chosen randomly and are thus often easy to weed out.
So there's really only a weak correlation between either of those metrics and real world performance.
There's absolutely a huge gap between user preference and model performanc that is widening by the minute. The more performant these models get, the more individual and syntactical preferences prevail.
> The most frequent failure mode among human participants is the inability to find a correct solution Typically, human participants have a clear sense of whether they solved a problem correctly. In contrast, all evaluated LLMs consistently claimed to have solved the problems.
This is exactly the problem that needs to be solved. The yes-man nature of LLMs is the biggest inhibitor to progress, as a model that cannot self evaluate well cannot learn.
If we solve this though, combined with reasoning, I feel somewhat confident we will be able to achieve “AGI,” at least over text-accessible domains.
I evaluated a lot of Math Olympiads in Argentina. Humans participants many times don't have a clear sense of whether they solved a problem correctly. We get review request of the grades, sometimes it's a mistake but the grader, sometimes it's a tricky error that changes the result very little, sometimes the human participant made a chain of huge mistakes that make no sense. (Obviously, we give a polite reply in every case.)
I don’t know enough about the USPS to comment, but I just want to take a moment to hate on Amtrak.
I know there is a huge “we love trains!” crowd among urbanists but Amtrak takes longer and is FAR more expensive on most routes than planes. Ultimately going all-in on airports was probably the better play than high speed rail.
I commonly see people say stuff like “what if you could get from NYC to Miami in 2 hours?” That would require a train ~15x faster than our current highest speed rail service, but only a ~30% faster commercial plane.
One of the highest frequented routes on Amtrak is NYC <> DC. This usually puts you out 200-300$ unless you can book months in advance (which essentially would defeat the point of a train but I digress) and takes 3-4 hours. By comparison a plane will take about an hour and if I book one just two days out that’s only $200. If I have a week or two of notice it quickly drops below $100.
It seems to me that people hate planes mostly because of airports and schedules not being convenient. This is a totally different problem than method of transportation, and one that is far easier to solve.
It would be infinitely easier and cheaper to make airports fast and efficient compared to expanding and maintaining high speed rail across the US. It is not fair at all to put USPS and Amtrak in the same bucket as while Amtrak is subsidized by the government it is insanely expensive. Americans are essentially paying Amtrak for the right to charge us ungodly prices in the name of “trains are cool.”
> I know there is a huge “we love trains!” crowd among urbanists but Amtrak takes longer and is FAR more expensive on most routes than planes.
This is not an Amtrak problem. That's a well known trait of conventional trains.
With high-speed trains running at 300km/h, there's a well known tradeoff point where high-speed rail is a far better option; up to 600km-800km. Beyond that, airplanes are still the preferable option.
The higher the train speed, the longer you need to travel for air travel to become the preferable option.
reply