113 karma · joined June 5, 2024
But I wouldn't use this often. I'd be willing to bet that you'd lose all performance benefits of using Rust versus something like Python or Ruby that uses forking extensively for parallelism.
I personally fucking love rockets and what he did. But for the past years made me hate the man.
My guess is that most people here are center-right wing, or at least a majority of HN isn't left-leaning.
It's surprising he managed to make even that audience hate him
It's the best parts of Tesla mixed with the best qualities of Korean manufacturing.
Volkswagen's ID5s are also very good since last year. Many of the software woes seems to be fixed.
I live in Europe and people absolutely went batshit crazy about it, not only in Germany. People really despise that gesture.
I'm willing to bet sales in Europe for Tesla FY2025 will be terrible.
I know many Tesla owners here that are considering selling the car because they are also afraid of the car dropping even further in price because too many people will want to sell it.
That's what being a leader means, you deal with the ambiguity, is paid more, but if things don't go as expected, you are axed.
The only thing that can save you is if you have built relationships with senior directors that could save you.
Just a reminder that Tomorrow the CEO can wake up and desire to cut people to increase their margins and Staff engineers working on improvements are the first to go.
There's no such thing as a career. Just focus on making money while you can.
Also, make sure you have a few doors open in case you need to get out.
This means you want to have a flexible skill set in case you need a new job, also a network of people that wants to work with you.
US companies will want to have consumers and Europe has them, so they would like to offer their services in the EU.
AI hardware is also produced in China. Americans don't want to talk about it, but Huawei has good inference hardware and software.
It will show you the maths, you'll build simple neural nets (from maths!) that can read digits and scale it from there.
While doing it, you may struggle with some of the maths, just take a deep breath and invest sometime to fill the gaps you need and continue.
Learn about CNNs and everything else step-by-step. It's awesome.
First, you build something that fixes a customer problem that they are willing to pay for.
Whether it uses AI or not, the customer doesn't care.
The only thing that matters is the value you add and your ability to sell and showcase how your solution is better and attract customers.
It just became really important in the last decade or so with Big Tech using them to filter candidates as a proxy for intelligence and dedication.
It makes sense for Big Tech to use it because they are huge companies, and often enough their engineering talent will work in very specific projects, rarely building features using a popular framework or OSS tools.
90%+ of the companies still don't use that method because they hire people that can use popular OSS tools to build software for them.
So they look instead to judge your experience in a certain framework, language or infrastructure piece.
So overall, I believe nothing has changed.
Mathematicians will still need to learn to solve integrals and remember rules without a calculator to pass their tests.
Coders that want to join Big Tech will still often have to do leetcode challenges without AI until they find a better way to filter for intelligence and drive.
You'll quickly notice how limited AI is.
It can generate some code, but often it's just boilerplate similar to what Frameworks and Libraries do.
It is far from fixing tough bugs, fixing performance or architectural issues.
Focus on what you think the AI isn't good at doing.
We don't write software, we value add to business
Now Meta will continue to destroy the world's mental health until the situation worsens more.
Trump was a hope that this Meta non-sense would end, but both became allies and forgot about the so many studies that shows how it's a toxic digital product.
It will now take longer for humanity to understand that Meta platforms are worse than cigarettes, but doesn't come with any labelling.
Meanwhile the US life expectancy has been decreasing.
And countries with huge governments like France have been doing just fine.
Also running a similar deficit than US.
Even being libertarian, I have common sense to acknowledge that those takes are plain stupid.
It's all about how the money is spent, not if it's public or private.
I bet most money spent by the government in the US is in entitlements, not infra or any kind of subsidies to key industries. So yeah, very inefficient!
I bet the EU and other countries will start giving them more fines.
Would be nice to see some competition though, that isn't either Tiktok or Instagram.
Soon we will get to know the consequences.
I ended up moving to Europe and working for lower tier companies and now at 36, get laid off left and right as the economy here sucks + my bad luck :-(
They fine tuned it for that test.
What we are seeing is a marketing trick to keep markets and investors excited about AI. It's a trillion dollar industry for NVidia and other players. Fake it until you make it.
If you look deeper, there's very little change since GPT-3.5 and Anthropic have catched up with everything OpenAI has built so far.
Sora was a huge fluke with other companies clearly ahead of it. Also mostly useless.
The numbers don't add up.
My whole life I've worked on places that work performance was everything in software engineering, many toxic places etc.
I hope at some point I can chill down and make peace with myself like you did. Congratulations!
I can already work extremely fast in Git by using bash aliases and my workflow.
This might be useful for those starting out.
Even if they use queues, I'm sure they are running at a loss and the GPU time is going to cost 100x more than what they charge.
Creating false demand for AI can easily bankrupt their business, as they will believe people actually want to use that crap for that purpose.
They also avoid doing fringe research in academia, making them not typically stand out when they happen to have researched the new hot topic.
I've known also a few people that either founded $100B+ public companies or did research like people like to talk about in HN "All you need is attention" and they weren't the smartest, but smart enough, but were quite lucky on their journeys and to be in the right place and in the right time
We just had a rate cut and in the past 30 years a recession happened after max. 2 quarters after the that.
If you see news from 2006 you will see people saying that the economy looks strong despite some structural issues, the economy is a huge machine and takes time to both take speed and stop, like a giant car.
Imagine the FED raising rates as pushing the brakes, the car won't stop suddenly, it needs time. The same happens with economic activity.
Now that the FED had to almost call an urgent meeting to cut .5, it's a clear sign that it could be that the car/economy was actually stopped. It isn't clear because everybody is working, but every cent that goes out of a central bank has an expectation of returns, if the expectations are that there will be negative returns, money will stop, it won't be invested, products won't get sold, houses won't be bought, everyone's plan will be delayed to the future.
And on Today's economy, you can't just stop or delay. Things need to be bought. A construction company needs to sell their real estate stock to pay back the bank and so on.
Money need to change hands, if it changes hands too slowly you get a recession and negative economic growth, if it changes too quickly you generate inflation.
I doubt we'll have a big bubble pop though, big tech will continue to invest in AI at similar values or increase the rate, what may change might be the hardware sales (bear on NVidia), but the run after AI talent will continue to happen for many years and possibly increase. Also, with more specialized hardware people might run away from expensive h100s towards cheaper alternatives or ASIC solutions.
Many small startups without a PMF will all fall apart, though, with OpenAI and other more established players surviving and potentially focusing on more profitable products (like ChatGPT) than newer initiatives or "AGI" research.
We've had upgrades to hardware, mostly led by NVidia, that made it possible.
New LLMs don't even rely that much on that aforementioned older architecture, right now it's mostly about compute and the quality of data.
I remember seeing some graphs that shows that the whole "learning" phenomena that we see with neural nets is mostly about compute and quality of data, the model and optimizations just being the cherry on the cake.
But some people like exciting things, to dedicate themselves fully to something, even if that something may yield nothing or even a very negative outcome.
Just see how many people die every year doing free solo climbing (climbing without ropes), yet there are still solo climbers. Often, they make 0 money and do this out of pleasure , fun and for being challenged and love of the sport.
Just because you like optimizing your life for the best outcomes (which typically involve less risk), it doesn't mean others will. Some people are just born differently.
Life is there to be lived. I don't work anymore ever as a founding engineer or founder as I've experienced loss and saw it isn't for me, but what happens is that you'll likely end up working for somebody that did that if you don't.
Even the founding engineer gets 10% or 1%, there'll be always the ones that hit the jackpot and build another Amazon, and even though they aren't rich like Jeff, they carry out immense power and got their billions. And they deserve it, because they risked a lot for that.
Being the unlucky one in this isn't fun. I've been there, saw it and it's ugly. But that's just how the universe is. Some will be lucky, that's something I'm sure.
I've already fixed a couple of tests like this, where people clearly used AI and didn't think about it, when in reality it was testing something wrong.
Not to mention the rest of the technical debt added... looking at productivity in software development by amount of tasks is so wrong.