cbg0

Nothing's stopping them from doing that, but you're incredibly naive if you think they're "protecting the average user" out of the goodness of their heart.

I've had no problems like the ones you've mentioned while using Opus 4.8. It does overthink stuff with higher effort levels but that's kind of expected.

This was attempted before in 2022 but fell apart because plaintiff couldn't show an agreement took place https://cdn.ca9.uscourts.gov/datastore/opinions/2022/03/07/2...

It's fairly obvious he's referring to 6.6.69, and I'd expect an LLM to make that assumption as well, unless prompted to be a pedantic HN commenter.

It will reduce usage and spur investments into competitors from the EU and likely other large nations, most likely hurting these upcoming IPOs.

You can negotiate your salary even without a union. Also being part of a union doesn't guarantee you won't be laid off because of AI.

How will the "feds" pay you a visit in Albania or China?

There is an example of this and you can probably dig deeper if you're truly curious instead of just assuming things.

https://www.wsj.com/world/middle-east/u-s-forces-raid-ship-s...

They absolutely can impose sanctions on foreign companies by restricting their access to US markets, investments and penalizing US banks doing business with them.

For some EU companies this is irrelevant, but for global companies this becomes a problem.

The US can create sanctions with legal repercussions, the same way they've done with Iran or Cuba in the past.

If the information isn't there why would they need safeguards against it?

I've played with smaller unrestricted local models and they will tell you how to make a bomb with easily available items as well as where to source them. I don't doubt that these >1000B frontier models have better information.

Just build a bot to bypass their bot.

I think it's based on a few sources + page 6 from here https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publ...

According to ChatGPT the labor cost per vehicle is more than twice as high in the US versus China. $1,341/vehicle vs $585/vehicle

https://chatgpt.com/s/t_6a292f0205ec819182b54e48ce9702a3

Solar panels, inverters and batteries are not critical infrastructure and I'd wager the jobs impacted are considerably lower than the automotive industry.

China, being a superpower, has a vested interest in bringing other superpowers down a peg as well as increasing other countries' dependence on China.

Their state has a serious incentive to ship out cheap cars to destroy the automotive industry in the US and the EU for example. When that happens they can double the price on all their vehicles and you can't restart your car factories to compete with them again until years down the line.

With Huawei its about telecom equipment which is essential in today's age, with TikTok it's about controlling the narrative, also essential.

Yes, US and EU manufacturers need to innovate and be less greedy but the cost to make things will always be higher than in China so even though protectionism sounds bad, you'll always have some of that around to even the playing field.

> There is no such thing as technical debt in the age of AI.

Producing new technical debt quickly doesn't mean there's no more technical debt.

50 billion dollars a year is never "almost nothing".

This comment smells like Claude.

You're making a bold assumption that this is the first time they're making money off of these bets.

The Q4 quantization requires about 600GB of RAM without context, not exactly consumer hardware friendly.

I haven't implied an AI takeover, this data will be repackaged into a product for military/intelligence, political applications, insurance companies that can charge you more because they know you're willing to pay, and many more.

These things already exist and happen, it's about the data getting better and not having to build tools to query it and make projections, since you can just type a query into a box even if you're not a data scientist.

> OpenAI did this with your health data in January. Now it wants your financial data too.

This is far more valuable, they can see what political affiliation you have based on your campaign donations, predict things like cheating on your wife & the impending divorce, what vices you have and they can also build shadow profiles of all the people you give and receive money from even if they don't use the product.

They gave away $100M in credits specifically for Mythos.

They announced the pricing when they released preview: $25/$125 per million input/output tokens. I have no doubt they're already selling it to select customers.

If only we had this tech back when managers were looking at how many lines of code you were committing weekly as a performance metric.

"essentially useless" is a gross overstatement. Your personal benchmarks will always provide you with the most value, but disregarding standardized benchmarks because you care more about vibes is not exactly scientific.

Your n=1 might not be very relevant outside your personal use. In less contaminated benchmarks Gemma 4 is way below Sonnet 4.5, let alone Opus models: https://swe-rebench.com/

The purpose is to control the total amount of requests they need to handle in a given timeframe. If everyone could use up their whole weekly limit in 5 hours, many would do so, thus pushing the GPU/TPU clusters to or above their capacity limits.

They're doubling the five hour limits, but no mention about the weekly limit. So overall it's the same maximum usage, right?