A lot of the insights of math come from knowing how to do things efficiently. That’s why the tests are timed. I don’t know, this is pretty basic pedagogy that you are choosing to grief.
It is a program. I need it to get task X done and I don't care how, whether it is strictly through CoT or with tools. There is no such thing as cheating in real work and no reason to handicap it. Just test the limits of what it can do with whatever means possible.
Trying to solve everything with CoT alone without utilising tools seems futile.
It’s not that bad. I’ve been using 3 Pro for some time now and I’m quite happy with how it works. Best paired with Opus and Codex, like most models, but it’s solid as a full-stack buddy.
I really want to like the native Mac app aesthetic but I kinda hate it. It screams minimalist but also clearly tells me it’s not meant for a power user. That ruggedness and sensitivity is missing.
Waymo needs $16B to build what Tesla already has: manufacturing capacity. Without that, there are only so many cars they can put on the road. They've proven they can do the rides. But they haven't proven they can do it cost effectively. To scale up and start making a profit, they'll need to start building/buying lots of Waymo cars. That's not going to be cheap or fast. That's going to involve a lot of capital expenses.
Tesla is the other way around. They can definitely make lots of cars and make a profit. But they haven't quite gotten FSD to the stage where it can do rides properly. Supposing they at some point figure that one out, they are very well positioned to start producing vehicles by the hundreds of thousands pretty soon after. That's indeed the premise for their valuation. It's risky but not completely without merit.
Another point to make is that Waymo and Tesla are not going to have this market to themselves for very long. There are quite a few autonomous ride hailing companies serving rides at this point. And while the attention is often on the US, China is moving pretty quickly as well. Several companies competing there in several huge Chinese cities, for example.
On the US side, I think there are a few players that might become competitive soon. Zoox is looking pretty solid. And Rivian is rumored to be pushing autonomy as well. There are a few more players in various stages of technical readiness.
The real battle will be in a few years when we are past the basic "does it work", "is it safe" questions and legal approvals all over the world become more routine. Then it will be all about volume and scaling. That's going to take probably at least until 2030.
Based on the news, I think Waymo will import base vehicle builds from China and then adding the control systems and software to those. So it’s not like they will start making cars.
That sounds right. Unless Waymo considers car manufacturers to be its competition and therefore something to commoditize, it wouldn't make sense for them to get involved in ground-up manufacturing.
And by this point, it seems like an electric-car platform already is close to a commodity, which is another reason for Waymo not to waste capital building another.
They probably see the electronics and software as their product, hopefully they will license it to someone so my next car will have it :). Lidar prices are cheap enough these days (but coming out of China, so who knows with Trump if that will apply to us).
Another approach is Waymo acquires Tesla's auto technology. Tesla sloughs off its dinosaur car business to focus on its new robot mission, and Waymo detoxifies the Tesla auto brand.
Aside from destroying about $1.25 trillion of market cap, this would leave everyone better off.
You don't really maintain electronics, you swap them out when they are detected to be bad. Electric vehicles don't need tune ups or overhauls, it is light maintenance and full on component swaps. Send the defective components back to the factory for refurbishment and/or recycling.
Nah. The local car dealerships and Tesla service centers seem to be pretty busy doing heavy maintenance on electric vehicles. The drive train might be marginally simpler but there are still a lot of moving parts that break or corrode just like any other vehicle.
I don't think $50k LIDAR sensor suites are disposable like that and the computer integrations is pretty sophisticated I'd imagine. These cars will take a beating.
Waymo has 0 need for manufacturing capacity. There are dozens of companies that do that really well at a low margin already that'll be happy for the business. They made a timing mistake by choosing Zeekr for it, which is limiting their expansion at the moment. That's a lot easier/cheaper/quicker to fix by choosing a different partner than by building their own.
Making a driverless car "driver" is clearly much harder than manufacturing cars though. Many companies manufacture cars and have done for decades. On the other hand Waymo is the only company that has actual driverless cars on the road. It took them a very long time. Tesla have been trying for a very long time too and still have a long way to go.
So IMO Waymo has something far more valuable than Tesla.
(Obviously the market isn't rational though so I wouldn't necessarily invest based on that.)
I don’t even think that’s rational, but it may be what’s propping them up.
Last earnings call Musk said Optimus wasn’t doing “meaningful work” at Tesla and as far as I’m aware they haven’t done meaningful work anywhere. I think they’re behind the curve there. Figure AI recently finished an apparently successful feasibility trial of their humanoid robots with BMW and Boston Dynamics has a deal with Hyundai for their Atlas humanoid robots.
I’m not even convinced humanoid robots are going to pan out in general. They only really make sense in a scenario where you’re back porting robotics to factories built for humans. That has value but feels temporary; factories designed to be robotic feel like the future, and there’s no need for them to do the job the same way a human would.
>I’m not even convinced humanoid robots are going to pan out in general.
I want one personally, so it can rake the leaves, mow the lawn, tend the garden, do the laundry and dishes, replace the roof, etc., when I'm old. But they should also be used to pick up litter along the highway, paint over graffiti, etc..
I feel like the humanoid form is getting in the way for that, and that a "Spot" like design with a hand on top is better suited for that. Also i think laundry and dishes are already 95% automated since about 50 years.
this is something that also never made sense to me - it felt like star wars got it right - for repairs and remedial tasks a trash can (rs-d2) or all the little service droids are more appropriate, but c3p0 or other nurse and protocol droids makes sense to look more humanistic since they serve functions to facilitate human activitiy - but there is no way those functions are numerous enough to be priofitable.
I _think_ these are meant to replace humans working alongside the industrial robots rather than the big industrial robots themselves. I don’t work in manufacturing though, and the press releases are too buzzword-y for me to grasp the actual tasks they’re going to do.
I would guess the long term strategy is to do this for economies of scale and then push into new markets opened up by the lower price point. I would guess these are horribly expensive right now, given something like Spot is way simpler and still like $40k
Probably because Tesla sells about a million cars a year, including the worlds best selling car (Model Y) since 2023. The stock consistently performs well as well, I know they outperformed estimates for last quarter. Being positioned well for autonomous driving presumably helps hold the stock up, but I don't think that's the core of the valuation, and Waymo does a fraction of what Tesla does. Waymo is impressive, but their 2025 revenue was ~350 million.
Race to the bottom, eh? Why talk about what the situation is in our country and try to improve it when other people in other countries have it much worse?
Oh, you don't have to out yourself like that; not here in public! Many people care about black lives and DEI. In fact, I'm willing to bet you probably agree with the most palatable form of DEI - jobs programs and hiring incentives for veterans.
In any case, here's a quote FTA:
>Rather than explicit imprisonment, the compound relied on a system of indentured servitude and debt to control its workers.
The catch-22 is that these people are nearly always immigrants, and the criminals have taken their documentation, so the best case scenario is they get rescued and then deported, possibly via a spell in immigration detention. The worst case scenario is the cops turn up, laugh, collect the day's bribe money and then the person who called the cops gets beaten.
(this is an important dynamic in sex trafficking as well)
this is far bigger a problem and requires interventions from China and India. what good is it to just punish the people who ran the scam but not the country that supported it?
>The relative leniency of Muzahir’s compound, says Harvard’s Sims, likely stems from scam operations’ sense of total control in Laos’ Golden Triangle region—a zone of the country controlled largely by Chinese business interests that has become a host to crimes ranging from narcotics and organ sales to illegal wildlife trafficking. Even human trafficking victims who escape from a compound there, Sims points out, can be tracked down relatively easily thanks to Chinese organized crime’s influence over local law enforcement. “These guys don’t have to be held in a cell,” Sims says. “The whole place is a closed circuit.”
But I think you are missing the bigger picture -- its true that LLM's might have a tendency to average out things but even if this is true, everyone needs to keep up.
People who use AI correctly will use it where they would have otherwise produced below average works.
There's many things to be unpacked here, but one that feels particularly compelling is how the potential of AI is being leveraged to divide people further along class lines instead of being a tide that rises all boats. The narrative around AI adoption is largely a rehash of "think of how much more we could do with less", which is generally great for the ones who owns the means of production, not so much for the ones who are on the factory floor. It's quite visible when AI is being used as an excuse to lay people off and expand the expectation on whoever is left without necessarily increasing compensation to follow with the "added productivity". The worker/employee power dynamic is only made worse there since it seems to be a shared corporate wet dream to hire as little as possible and have "digital coworkers" fill the gaps. This isn't a new story: before, it was offshoring to cheaper markets (i.e. where the power dynamics are even more tipped toward the rich), now it's the dream of a digital worker that doesn't talk back and will work 24/7.
There doesn't seem to be much interest in using this to make people work less and engage in the rest of life more, hence the jokes about how somehow, the machines are generating art while folks are still toiling away to make ends meet.
The class division is even in the original message we're responding to: failing to get on the bandwagon or being pushed from it leads to lower income and status despite there being plenty to go around if we only stopped chasing infinite growth.
The only argument I've heard to far about how this is the path to betterment relies on getting past a theoretical hump where AI and robotics go further and can replace most labour, and the upper class giving up it's power over the masses voluntarily instead of going further down the same road we're already on.
I can't speak much for the first, but I think trickle-down economics has better chances of working out than the second (i.e. none).
AI can help increase productivity, that is get more outputs (goods/services) for the same inputs.
Two hundred years ago, shirts had to be hand-spun, hand-woven, and hand-sewn, so ordinary people could only own one or two. Now because of automation and factories they are so cheap that poor people have many.
Previously 97% of the workforce was engaged in agriculture, and even in the 20th century, famines killed millions. Now with increased productivity we create so much food that obesity is the defining health crisis of our time.
All classes have been able afford better clothing and more food, not just those owning the means of production.
> instead of going further down the same road we're already on
Even in the last 25 years, we've seen large increases in life expectancy, child mortality fell by more than 50%, 1 billion people left extreme poverty, access to knowledge and education expanded, and more.
I don't buy too much into your framing of pitting classes against each other.
But I think we both can at least agree on this - AI can actually replace some labour or augment it enough to require some shuffling.
My argument is that if AI brings efficiencies, products become cheaper and cost of living reduces. It's how it always worked - agricultural revolution, industrial revolution and information revolution. Cost of living reduced constantly because efficiency increased.
> upper class giving up it's power over the masses voluntarily
The upper class needs labour to consume their products. So it is in both interests that both labour has more money to spend. What then happens is, people are richer but the upper class gets even more wealth. Everyone benefits.
How you work without calculators is a proxy for real world competency.
reply