anandnair

We don't even need that example. The example is in front of us. Take a smaller parameter model and ask it to do the same complex thing that a larger parameter model did. It will struggle.

Btw, I'm not saying it's just the number of parameters that matters.

I think we're drastically oversimplifying what "pattern matching" means. It is also one of the fundamental mechanisms by which the human brain operates. I believe we are consciously (or perhaps subconsciously) conditioned to think that human "logic" and "reasoning" are several degrees more advanced than pattern matching. However, I don't think this is true.

The fundamental difference lies in how patterns are formed in each case. For LLMs, all they know are the patterns they observe in "words" - that is the only "sense" they possess. But for humans, pattern recognition involves continuously ingesting and identifying patterns across our five primary senses—not just separately, but simultaneously.

For example, when an LLM describes something as "ball-shaped," it cannot feel the shape of a ball because it lacks another sense to associate with the word "ball-shaped." In contrast, humans have the sense of touch, allowing them to associate the word or sound pattern "ball" with the physical sensation of holding a ball.

Coding, especially the type mentioned in the article (building an app based on a specification)—is a highly complex task. It cannot be completed with a single prompt and an immediate, flawless result.

This is why even most software projects (built by humans) go through multiple iterations before they work perfectly.

We should consider a few things before asking, "Can AI code like humans?":

- How did AI learn to code? What structured curriculum was used?

- Did AI receive mentoring from an experienced senior who has solved real-life issues that the AI hasn't encountered yet?

- Did the AI learn through hands-on coding or just by reading Stack Overflow?

If we want to model AI as being on par with (or even superior to) human intelligence, don’t we at least need to consider how humans learn these complex skills?

Right now, it's akin to giving a human thousands of coding books to "read" and "understand," but offering no opportunity to test their programs on a computer. That’s essentially what's happening!

Without doing that, I don't think we'll ever be able to determine whether the limitation of current AI is due to its "low intelligence" or because it hasn’t been given a proper opportunity to learn.

Curious to know how you'll manage the infra cost if customers just pay $15 for lifetime

Haha LLMs are trained to be obedient, secure, and trustworthy, until you ask nicely :P

If I don't enable cloud sync, would it be stored in my localstorage?

Why do you think Google went wrong?

Cool. Does it depend on external tools, for example to compress PNG, change GIF speed etc?

This is actually a great idea. I haven't used it to know how perfect the photo will be. The only concern I have around is privacy. Maybe you should mention that the photos get deleted from your servers after we download it.

Cool project. How does this compare to Cursor composer? I was planning to get the pro subscription for Cursor :)

There is a change in trend of-course. For example, before AI we used to hire freelance devs to build small, one time projects, or build/maintain client libraries on top of our REST APIs. But now, we can just do it ourselves! So, The impact is true.

I don't think building custom solutions will give the best ROI in the long run. In JS ecosystem, I rely on npm vulnerability warnings to figure out if something is broken.

Website is not working?

Congrats on the launch. This looks cool.

This looks cool. My concern about adding such a middleware to the API routes is the potential performance hit. Assuming this middleware would call your APIs to track metrics, wouldn't it degrade the performance of the existing system by consuming some CPU power?

I think my question wasn't properly understood. I was thinking whether LLMs should be considered as a "fresh new" smart brain that can learn anything, just like a smart human kid.

Or should it be like a brain that comes with preloaded knowledge? (like how it is today)

Not available for Mac?

Dev tools market is indeed an interesting business to make tons of money. Because of the huge value that it offers to businesses. If you look at the dev tool product companies (even open-source ones) started 10 years back, most of them were either acquired for billions, and many of them did IPO. And most VCs missed this opportunity since they were skeptical about this space.

I think that is what is driving the current VC funding trend in dev tools. Also, considering the future of software engineering, dev tools will have even more adoption than before.

And regarding the pessimistic comments in HN, they are mostly coming from people who lack in-depth knowledge about how businesses work.

Why should people choose this instead of the competitor if it's a copy?

A CTO should be the bridge between technology and the business. They should understand both business and technology. You'll learn things as you grow. But the CTO mindset is important. Your job is NOT JUST to be a developer, an engineer, or an architect. The decisions you take can have a huge impact on your business, and the bottom line. Your experience at FAANG definitely helps. But think beyond that engineering experience. You're building a business. Can you foresee where your company is heading? Can you recruit, train, and grow a team of engineering leaders who will also steer the company in the same direction?

From day one, keep these in mind

1. Anyone can build anything, given they have the time and money, and resources. But only a few people can understand what their customers really want, and build that.

2. You are not an engineer. You are an entrepreneur.

3. You are not here for a solo game. Can you be a great leader?

4. You should understand how your decisions affect the company's finances.

I just received a notification for this comment. Thanks to my extension ;) I think we lose the opportunity to engage in meaningful conversations if we don't get notified effectively. (Well on the other hand, too many notifications make it counterproductive though)

SSL Error on your website.

Look at Apache Airflow

I'd say, development is much faster using GraphQL than REST.

I don't think there is a foolproof way to tackle this.

Yea, technically it gets lighter because when you're writing a bit of data, you're storing a positive charge, which means an electron is removed. But what about magnetic storage medias?

apparently, lol

[Update] - I ran the code again with proper compiler optimizations and then C++ performs slightly faster than Rust, thus making it the fastest among all.

oops, edited

I did it, and now c++ is slightly faster than Rust :)