iamandoni

This is exactly why I built https://github.com/andonimichael/arxitect . I’ve found that agents by default produce tactical but brittle software. But if you teach agents to prioritize software architecture and design patterns, their code structure becomes much much better. Additionally, better structured code becomes more token efficient, requires less context to make changes, and coding agents become more accurate.

Larger composition, though, starts to run into typical software design problems

I've been seeing the same thing. Where agents are great solving the immediate task, but as changes compound they run into software & architectural design problems. I created https://github.com/andonimichael/arxitect to help at least have coding agents self reflect on their software design. But I really like your approach to self-modification and improving the agent itself instead of just teaching it another skill in it's context.

I can’t imagine that will stay the case though. I built https://github.com/andonimichael/arxitect as a first step to using agentic coding in a more production ecosystem. Agents will be able to write useful (and high quality) software over time, their training has just under-prioritized code quality thus far

Do you take any measures to prevent link hallucination? And content grounding / attribution verification?

Shadcn & Vercel have created https://v0.dev/ which is pretty incredible

Check out the Sniper in Mahwah blog. Most of this space is proprietary, but the author uncovers a lot of the landscape from public records.

An alternative I’ve been enjoying is @tommertom’s Ionic port (+ Capacitor) for native integration: https://github.com/Tommertom/svelte-ionic-app

Wow what a great tool! Intuitive and clean interface with realistic default equity investments/option pools. Also I love to see more and more svelte projects popping up. Really exciting seeing its growing adoption :)

I believe there is a fallacy in the analysis of SF vs neighboring counties. The author is comparing YoY change; however, if the adjacent counties have less baseline nominal crime the YoY change could be less while hiding the magnitude of crime in SF. The author also doesn’t differentiate the types of crime in the analysis which makes a substantial difference. This section is in reference to the lenient theft laws, whereas these numbers commingle that data with violent crime and all other crime. And lastly, COVID is an outlier year where the population of the city dwindled, there were massive store closures, & there was less opportunity for non violent crimes.

The book Faking Science by Diederik Stapel is an excellent view behind the scenes of a serial academic fraudster.

I would advise you to think about how to make that value proposition proprietary and defensible. I’m not wholly convinced that offering interior design or offering whole-apartment bundles alone is something you could defend an incumbent from also providing. Although it does sound like your space-visualization technology might be a leg-up here.

Also note that there are several competitors from the other-side that provide whole furnished apartments for rent such as Blueground. You seem to be in between the “rent single pieces of furniture” and “rent a whole apartment” spaces and need to ensure that the space is carved out enough and the market is there.

Congrats for being accepted into YC! You now have a large journey ahead of you!

I thought he brought up really good points about the politicization of current social networks. I miss the days of chronological feeds and only seeing posts of users in your circle. I think splitting up actual users and pages/groups you follow into two separate feeds will also help reduce the feedback loop Cuban talks about.

While I mostly agree, getters without setters is a nice way to expose that youre data is immutable rather than relying on final fields

Everyone is bashing this as if it's a peer-reviewed paper...Note guys: This is just two students' final project for a class, not some security experts' publication

Adleman, Rivest, and Shamir introduce public-key codes, a system for passing secret messages using large primes and a key which can be published.

My OCD wishes they would have said Rivest, Shamir, and Adleman like the algorithm is named after

That invisible ink at the top is a good subtle touch.

If you liked that, you should also read Kevin Poulsen's Kingpin. Absolutely enticing story.

This is a pretty cool browser game that uses device orientation and websockets: https://lightsaber.withgoogle.com

I do not know about inviting - as I was only invited in the US. However, when they asked for which office I wanted to interview, I chose from the global list of offices.

As a correlation to this site, there is a secret programming challenge within google at: http://www.google.com/foobar/

You have to get invited to it (via searching the correct keywords, etc. I got invited when googling about dependency injection). It is a series of programming challenges that get really difficult. But after completing them, I got a final round interview at Google, which was pretty cool.

I don't understand why there is a controversy at all here. This seems to at least have a logical use case, which is more than we can say for status code 418 (https://tools.ietf.org/html/rfc2324)

I don't know if you're actually interested in the theory behind analyzing game complexity or just trying to be negative without doing your homework.

Classifying traversal and 2-D games is not a new thing, and they simply apply previous methodologies and proofs to classic Nintendo games. They do not do this via "proofs by obscurity" but rather by classical NP-Hard proofs built upon previous theorems. Specifically, I will point you to the following two papers on fundamentals of proving NP-Hardness of 2-D and traversal games respectively:

- http://link.springer.com/article/10.1007%2Fs00224-013-9497-5

- http://link.springer.com/chapter/10.1007%2F978-3-642-13122-6...

They chose the collection of games because of their relevance to the authors. They wanted to study classic Nintendo games. Other games have also been studied including some, if not all, of the games you mentioned. If you did your research, you would already know the answers to the complexity of games like Tetris or Pong.

If you're genuinely interested in the underlying proofs or the theory behind game complexity, I urge you to read through the papers and their citations first before jumping to conclusions.

From the paper:

For these games, we consider the decision problem of reachability: given a stage or dungeon, is it possible to reach the goal point t from the start point s? Our results apply to generalizations of the games where we only generalize the map size and leave all other mechanics of the games as they are in their original settings. Most of our NP-hardness proofs are by reduction from 3-SAT.