steno132

My first time using Grok. I'd been so used to using AI models that declined to do things I told them, like tagging people in a video feed, helping me "optimize" my taxes or managing my Twitter bot farm.

Grok just did these things for me, no questions asked, no ethical judgments. No woke.

Elon really doesn't get enough credit for Grok. People don't want the most powerful reasoning model or "constitutional AI". They just want a model that does what they say. Elon understood that insight (like he usually does) and no one else really did and that's probably why Grok has been growing rapidly over the last two years or so.

But practically AI inference requires substantial local computing resources. It's not some web app, it's a order of magnitude more compute needed

Agreed. I feel like Alphabet (who wrote this post) has kind of lost track of what they're espousing here.

Self driving cars were the "monkey" 10 years ago but now they are largely derisked. I'd like to see Alphabet working on the next generation of innovations. For me personally, I'd like to see them work on:

- Teleportation. - Talking animals. - A cure for death.

I'm not seeing them work on anything remotely as ambitious these days!

Seems like you tore through some outdated Jira tickets, which is cool, but what was the impact of your work?

Hard doesn't mean impactful. In fact that's one of the biggest mistakes I see junior engineers make! But glad you got that learning experience, as you become more senior you'll definitely develop a better understanding of what drives actual impact.

This is a great initiative. Postgres was written in the 1980s and we can't afford to have our most utilized workloads running on a software written before most of us even existed. LLMs make it possible to rewrite Postgres and we should take that chance.

What I'm saying is that Postgres was built for a long gone age. We need a extensible database written in Rust which can serve as a foundation for any data system. We don't need a relic of the 1980s serving our most critical workloads.

Take a example. You could be a hunter gatherer, you could grow all your own crops from scratch, you could forsake all modern medicine and rely on superstition - would you be considered a wise person? You figured out things a lot of people didn't, you took the harder road - but most people would say this approach is just dumb.

What's wise isn't forsaking technology, it's using technology to improve things, and developing new technology as well. What's wise is using AI as a power user and seeing how you can contribute to the intelligence age.

Celebrating a foregone past where human intelligence was more valuable is hopeless and naive at best.

My view is, if you're going to use the service - you should give the data.

It's like using Gmail and expecting them not to train their AI models on your data - how can you expect that when they're giving you a secure, reliable, highly functional email client completely for free?

The digital economy only works if everyone pays their fair share. If you don't want to give your data then you are really harming everyone by slowing down AI development for everyone else.

Exactly, thank you, we are on the same page! It's great to be able to use our own devices and not have their compute coopted by a third party.

I'd rather not have intensive compute needed shifted onto my personal machine which I want to use for something else.

What's wrong with the data being used for something else? Someone is providing digital intelligence to us, saving us many hours a week, so the least we can do is provide them a little data so they are able to improve their service.

It would be selfish and unethical not to in my view. And ultimately the data is just being used in order to improve the models and benefit us, not for anything nefarious.

GPT1 was released in 2018, so yes, since then.

Practically if you're running a small personal automation project you're not going to want to waste a lot of time configuring and tuning a local model. You want to build the automation and move on.

If you're building a automation as a company you definitely won't want to take on the long term maintenance overhead of running your own models for some automation project.

This is dumb. Don't do hard things, do things which have a high ROI.

I could try to swim across the Pacific Ocean, walk across Siberia or consume a family size portion of McDonalds in a single sitting. All those things are hard. All those things have zero to negative ROI.

Doing hard things won't get you anywhere. Most hard things are useless. And a lot of the most valuable things are not hard at all.

Test driven development is one of the worst ideas nowadays in the LLM age. We have models that can consistently write expert level, usually bug free code for you and rapidly fix even complex bugs in your codebase.

The token cost and tech debt introduced by tests is just not worth it. There's usually no bugs and if there are, you can fix them quickly if and when it's needed.

I don't get this obsession with smaller models. I've been using Claude and GPT models for years and have had zero issues with them.

I see absolutely no benefit to me as a end user for a local model which is going to take up more of my CPU and memory and slow down my machine. I almost always have Internet and if I don't then not having access to a AI model is the least of my concerns.

This is just narrow thinking. Say Claude did increase the bugs in rsync by a negligible factor.

So what? You've saved a significant amount of time for a decent number of humans, and if those humans are working on other projects, the overall net output for the world is net positive compared to without LLMs.

You have to broaden your perspective. It's not just about how rsync was affected.

I would argue that for all but the largest tech companies you only need a single data system which is Postgres. Message brokers, analytical databases all can be built on Postgres. Unfortunately, Postgres as it's built now lacks any semblence of extensibility which makes this impossible in practice.

I would propose a rewrite of Postgres in another language like Rust, introducing a pluggable application layer on top. While ambitious in scope I think it would be helpful and even necessary.

It's really hypocritical for the US to cooperate with Russia on space, even take their help on repairing the ISS, only to sanction them and even their trading partners for buying their oil.

It's a hot take but I do think the US should be more appreciative of Russia's longstanding contributions to the ISS and other space projects of international cooperation and factor that into sanctions decisions. We do need their help as much as they need ours in space, and the fact that they are still helping us despite our treatment of them speaks volumes about their leaders' character.

This is a horrible take if for no other reason that digital intelligence is one of humanity's greatest achievements. It took thousands of years of advancements across chemistry, physics, material science to reach the point in the tech tree we are today.

There's absolutely nothing wrong with using AI and it's something we should be celebrating.

Java developers tend to produce more value since they can write the code in a more concise way using Lambdas, functional interfaces, record classes, etc. Overengineering like in your example above can be done in any language, but shouldn't.

Yes, SIMD is possible. But what about the other more obscure hardware features there's no nice library for? It's much easier to use those in Rust than Go.

See, this is a problem I see many junior engineers struggle with: Excessive desire to make one's own life easier at the cost of customers.

Yes, your codebase is nice and elegant and maintainable. But you haven't shipped as many features as the Python or Java developer, haven't delivered as much value to customers. Ultimately, the users suffer. It's a form of selfishness.

Agreed. We should make life easier for our users. That's why a language like Python or Java, with a faster development time, where we can fix bugs faster and ship features faster, is vastly preferable.

Hot take: There is no good use case for Go.

For infrastructure projects like PeerDB: Use Rust. It's significantly more performant out of the box, and allows you to use low level primitives like SIMD to make your project even more performant, which isn't possible in Go.

For applications, use Java or Python. The dev productivity is much higher versus a verbose language like Go. That matters more than the higher raw performance you might get from Go (for this use case).

For all the hate: Elon ships. And OpenAI ships.

People claim OpenAI is closed, that they are controlled by Microsoft, that they don't care enough about safety...

But the fact is, Anthropic, Google Brain, even Meta -- OpenAI blows them all out of the water when it comes to shipping new innovations. Just like Twitter ships much more now with Elon, and how SpaceX ships much more than NASA and Blue Origin.

If you disagree, give me just one logical reason why. It's just a fact.

Elon’s arguably the greatest founder in history. Either him, or JP Morgan. He doesn’t need to be more like a physician, physicians need to be more like him.

I've said this before and I'll say this again: Space is not a amateur game.

I see a lot of young SV types thinking they want to build a deep tech space company. And they can. But it requires a couple years of deep training in of aeronautics and physics before you even know what's going on.

Elon for instance spent years studying space on his own before launching SpaceX. Elon's at a different level, but we can at least learn a few things from him and SpaceX.

I once said that "just as there's no founder like Elon, there's no Java expert like Andrew (project author)"

Andrew's Java skills are at a different level. I only briefly met him once, but he used to be a core contributor to Oracle's Java magazine. 80% of the "Unix greybeard" type advanced Java knowledge I have is due to his articles.

Happy to see he's still building. Thanks Andrew!

Hard disagree. What's the last major feature Twitter shipped before Elon? Keep in mind they had 8000 employees

Just in the last one year, we've gotten: Twitter Notes, long form video, checkmarks for everyone, subscriptions for creators, edit tweets, see # of replies and likes on a tweet

Yes, Twitter is worth less than he paid for it. He's partially responsible, he's been too erratic. But I'm optimistic with this shipping velocity, in a couple years the company will be worth much more.

Twitter Notes, long form video, checkmarks for everyone, subscriptions for creators, edit tweets, see # of replies and likes on a tweet

And that's just in 1 year