f0e4c2f7

This is an extremely good insight and piece of advice. Already we're starting to see what looks very similar to a "frontend" and "backend" kind of development model similar to the previous tech generation.

While there are advantages to going "full-stack" in this analogy most people focus on one or another

Is there a way to try this yet?

Lately I've been experimenting with using TDD approaches for LLM assisted coding. There is a little bit on YouTube and some research papers on this topic. But most of it relates to how well you understand the test suites for the given language / framework to ask for the tests to be designed in the right way.

This has been great in gpt-4o for python because it has a built in interpreter. So I can describe the way I want the API to look, ask it to write tests in Python, write the code to satisfy those tests, and then run the unit tests in it's own local python env to see if the code works.

This streamlines things a lot. I find I can get working artifacts from gpt this way when other I often have to have a bit of a back and forth where I'm the one testing and pasting in errors.

For gpt-4o this only works with python though (for now).

It might be interesting to build a platform like this, even in your local laptop that allows the LLM to shell out to a sandbox for something like React with the unit tests configured so the LLM can run them and get the output.

Interestingly, frameworks like react also have test suites around ui/ux concepts. I haven't tried it, but I think it would be interesting to use that same approach with those test frameworks to attempt to get better looking frontends (something the LLMs are not especially good at out of the box)

Exa is trying to do this. I've found some sort of interesting stuff this way but it honestly doesn't feel quite good enough yet to me.

https://exa.ai/search?c=all

Does this still work?

If you follow someone you're still more likely to see their posts. You just see other stuff too in the "For You". It's algorithmic, and at times annoying - but contrary to the article is still very valuable in my experience.

I would say twitter stays about 1.5 months ahead of every other site I know of on AI news in particular. I've usually had time to demo new tools for quite a while before hearing about them on HN or elsewhere.

Architect the program yourself in your head. Then use gpt to write one function at a time. For obscure libraries, paste in documents directly to the context window and give links to example GitHub repos. Paste errors in to troubleshoot as needed. Architect for many small libraries in their own files so it's easier to compartmentalize and let gpt work with smaller blocks of code.

This study uses a version of ChatGPT that is either 1 or 2 versions behind depending on the part of the study.

It cracks me up how consistent this is.

See post criticizing LLMs. Check if they're on the latest version (which is now free to boot!!).

Nope. Seemingly...never. To be fair, this is probably just an old study from before 4o came out. Even still. It's just not relevant anymore.

Lots of good tools for this.

One way to do it completely free is OBS + Kdenlive. The interface for both leaves something to be desired but both open source and have all the features you would want (though sometimes buried in menus)

I find that the key is in working in the right size abstraction. If you ask gpt-4 to write whole apps for you, it usually can't get there. But if you architect the app in your head you can ask it to build it out one function at a time (with unit tests) pretty effectively.

When I'm testing out a new LLM this is usually my go to. How far up the abstraction layers can I go?

Still have to check the code yourself, ask it to troubleshoot etc. but I find this workflow pretty effective.

GPT-4. I use this for learning things but also if there is a clickbait article with information I want to know, I paste in the entire article and ask my question. I used to just abandon clickbait and accept not finding out, now that never happens.

Ublock Origin. Assuming most already know about this one.

YouTube Premium. YouTube has some of the most amazing content on the internet currently. Pretty cheap to go adfree and you get unlimited music with it.

Sponsorblock Browser Extension. This skips sponsorship ads in YouTube videos. Like an adblocker for in video sponsorships.

Wikipedia Current Events. This page is one of my favorite less biased news sources.

I'd still like to do a lot more with this kind of stuff. Now with LLMs, I think in an ideal world I would like a browser on my phone and PC that runs all content through an LLM to automatically filter stupid bullshit. I think we're not that far from being able to build this today (maybe already).

Big tech.

Practice leetcode hards and read the books on teachyourselfcs.com to prepare for the interview

Startup time! Has its own tradeoffs but you get to go compete with the system that is maddening. Imagine them trying to run a race against people who are actually trying? That's why startups are able to win against huge companies with seemingly unlimited resources. They use those resources to fight internally over who controls the resources.

Startups have their own perils of course. But you get to do work. Real actual work.

Nah it's just as goofy now if not goofier. What's hard to remember is that back in 2013 nobody really knew or cared who Travis was yet. You can never predict the successes. I guarantee around that time there were much more "serious" people who looked more likely to succeed.

Likewise, now some people may seem really adept at startups. Perhaps they know all the right VCs, have a PhD, maybe they grew up in SF instead of moving there.

At the end of the day that stuff matters very little. The next big startup will be from someone, and something no one sees coming.

For example, I can think of one AI founder who is growing his company almost entirely on the back of using his product to make and post completely unhinged memes.

Harder to imagine in 2013, even then it might have seemed a little too unserious.

Poker. Game of skill. Looks like gambling from the outside. Rich people use it as a way to socialize while wasting money. Not dependent on any technology. Has existed for hundreds of years (Lindy).

Practice enough to win at an average table and then know at any time you can get on a plane for Vegas, Atlantic City, or another place with consistent games and make a living playing cards.

Perfect as a backup plan / hedge. Poker also teaches lots of life skills so still useful if you never really use it too.

This article feels fairly naive to me. One possibility is that the EU is doing what is written on the tin and executives and companies are foolishly misunderstanding.

Another possibility is that reality is more nuanced and that executives are playing edge cases where profitable, like they do in other countries and jurisdictions.

Instead of seeing the EU (or any other system) explicitly as what is described in founding documents or official meetings, there is another perspective.

https://en.m.wikipedia.org/wiki/The_purpose_of_a_system_is_w...

Nope. I use it more than Google now. Most amazing piece of software I've ever used by a country mile.

Long term, my plan is to buy a cheap EV like a Bolt as well as a powerwall or other home battery in a part of the world with good sunlight and build a nice little solar farm to power the house and car.

Solar panels just keep getting cheaper and this plan seems increasingly like a way to nearly do away with power bills all together while still staying on grid.

But until then I tend to agree, EVs aren't as appealing when you can't predict what energy prices will be over 10 years

For context, Elon's comp was also put to a vote in an annual meeting, passed, a minor shareholder (9 shares) filed a lawsuit and here we are.

In the case of Tesla in Delaware at least, the court did not optimize for shareholders wishes. This is what has Elon wanting to move out of Delaware as much as the comp.

AI probably feels crowded but it's really not. Still very early for mass adoption and even within tech it remains somewhat niche.

Seems like there is an opportunity here for a workflow involving clipboard management and GPT-4V or another vision model.

We're currently going through the largest tech downturn since the dot com bubble. It's not you, it's the market. Tough for everybody who is looking right now.

For the first time in 20 years there are more applicants than jobs again.

Saw this on Twitter. Search Engine where the results are created by Mistral LLM.

Fun to mess with

This is part of a longer series but is maybe the single best video I know of on the topic:

https://youtu.be/kCc8FmEb1nY?si=zmBleKwlpV06O3Mw

I thought this video from Steven Wolfram was also quite good:

https://www.youtube.com/live/flXrLGPY3SU?si=SrP1EJFMPJqVCFPL

I don't think this is the only factor but I suspect part of it is because GPT-4 had access to better datasets.

After ChatGPT came out a lot of the places GPT was assumed to be trained on (reddit, twitter) started closing their APIs.

This alone represents a pretty significant moat.

Targeted advertising creates an incentive to closely track and monitor users across the entire internet. This tracking data is then collated, packaged, and sold to various parties (advertisers, private intellegece companies, etc)

Without targeted advertising this cottage industry and the associated data on people would not have as much of a reason (or any) to exist, improving privacy for the average person.

The first 30 minutes of this interview does a great job of explaining what's going on here. Really interesting stuff.

https://youtu.be/qcvMjoJdck4?si=RL1WnAAj4loS1D2s

It's pretty wild how baked into modern life insecure 2fa is. Especially with the prevalence of sim swapping. I more or less model most auth as trivially insecure at this point.

You think about someone like Vitalik of all people, if he can't keep his account secure...average person has their work cut out for them.

Private key auth systems have security challenges of their own (losing access forever when you lose your key) but I wish they were an option in place of the current regime.

In the 90s you could bypass security locally on a machine by clicking cancel and it would just log you in. Feels like today it's only slightly more complicated and costs a bit of money to access twitter, email, bank accounts etc.

Seemingly little to no interest in resolving this state of affairs beyond obscure and increasingly less legal crypto based systems.

Glad to see this comment. Had the same experience a while back. I think it was with this person? I'm not sure. I don't really remember the name and they may have changed it since then.

The experience was enough to make me wonder if there actually even was a discord group or if this was some kind of a phishing or marketing scheme.

Lots of friendly communities out there on discord that you can join without running into whatever kind of status game or possibly scam this is.

Best of luck on your game!

Cool idea. One more fictional example:

https://www.youtube.com/watch?v=IWIusSdn1e4