alsodumb

I feel like a lot of issues you mention will eventually point to an ambiguous spec. I’ve had great success with interview style spec generation where I ask agent to grill me on any and all parts of the spec/plan that’s ambiguous and then note my responses. It takes like 10 mins but the end product is always exactly what I expect and most of the things you mention get resolved in this interview phase.

My hand writing is so bad I don't know if a really want a font out of it lol (love the project though!)

It takes hundreds of millions to build a warehouse. Amazon has tons of them. Retrofitting things is capital intensive.

I blame the transit agency for those missteps though - it doesn't have to be like that.

Take for example CUMTD (mtd.org), the transit agency serving Champaign-Urbana, a college town in Illinois with about 200k people. It's an excellent bus system, everyone in the city loves it, the people running the place always embrace new technology, and they actually have a hydrogen plant setup in their depot and the plant is powered 100% by solar energy: https://mtd.org/inside/projects/zero-emission-technology/

Honestly I don't think Anthropic cares about you moving over to them - it's pretty evident that they already have more demand than they can handle.

I've always had better experience with Claude in day-to-day coding and text writing, and looking at public forums that largely seems to be the case.

1X is different. This is purely a hardware demo, they are doing pure teleop on the software side, no AI in the demo (except maybe RL for walking). 1X's bit thing was always their mechanically compliant safe hardware.

It's teleop, they are pretty open about it. It's not autonomous.

They do have an RL controller running for the legs, but it's just an intermediate controller and it probably get high level commands from a teleoperator. The upper body is purely teleop.

When Valve came out with their VR headset that had base stations, everybody thought that’d be the holy grail, that you can never achieve better localization and tracking without base stations, and a base station free method can never be better than that.

Well, Meta poured a shit ton of money into making Quest base station free and they got there. We use to use valve setup for our robotics applications but we swapped it out with Quest cause honestly Quest was as good but much more easy to setup and operate.

The bitter lesson is that don’t bet against data or compute. Also, I don’t think you’d have to train a AI model for each location at every time in the future. Things get more efficient, etc.

Vanilla vscode has some version of this in the debugging mode, you'd perhaps need to install their Python extensions but that's about it.

Yes! Rerun can definitely make your life a lot easier!

Rerun natively supports the server and the viewer being on different devices (https://rerun.io/docs/reference/sdk/operating-modes). In your case, in the script you are dumping data into csv, I'd add the relevant lines to log data to rerun.

On the desktop side, you can spawn a viewer that can listen to the stream and visualize it.

I am not the hiring manager, and unfortunately a lot of interviews and hiring decisions happen at org level or at my manager level. These are mostly sophomores/juniors - folks who went through school in post-COVID, post-LLM era with a lot of virtual classes.

I tried my best to explain it to them, and nudge them to using docs. I did live debugging sessions with them to try and 'teach' them how to use docs. Ultimately, it was taking away too much of my time for little to no return. I only started working in the industry like a month ago and it's my first time having interns that I didn't pick (back in school, I had undergad research assistants that I interviewed/selected, and they were all excellent) - still learning the ropes.

I am a robotics engineer/scientist and I do shit ton of visualization of all kind of high-fidelity/high-rate data, often in a streaming setting - time series at a few thousand Hz, RGB/depth images from multiple cameras, debugging my models by visualizing many layer outputs, every augmentation, etc.

For a long time, I had my own observability suite - a messy library of python scripts that I use for visualizing data. I replaced all of them with rerun (https://rerun.io/) and if you are someone who think Scipton is exciting, you should def try rerun too!

I use cursor/vscode for my development and add a line or two to my usual workflows in python, and rerun pops up in it's own window. It's a simple pip installable library, and just works. It's open source, and the founders run a very active forum too.

Edit: One slightly related tid-bit that might be interesting to HN folks. rerun isn't that old, and is in active development, with some breaking changes and new features that come up every month. And it means that LLM are pretty bad at rerun code gen, beyond the simple boilerplate. Recently, it kind of made my life hell as all of my interns refuse to use docs and try using LLMs for rerun code generation and come to me with a messy code spaghetti. It's both sad and hilarious. To make my life easier, I asked rerun folks to create and host machine readable docs somewhere and they never got to it. So I just scrape their docs into a markdown file and ask my interns to paste the docs in their prompt before they query LLMs and it works like a charm now.

I agree that they weren't explicitly meant to represent geography, but implicitly they do, right? Are there cases where this is violated?

In other words, is it safe to assume that for entity in a zip code is less than x distance away from the closest entity in the same zip code?

Route 4 (the helicopter route along the river) has a maximum altitude of 200 feet for helicopter. From the crash it seemed like the helicopter pilot wasn't strictly following the ceiling.

Unfortunately it's not gonna happen. The traffic at DCA only keeps going up, and I've heard chatter that it's because a lot of congressman and senators prefer taking flights from DCA instead of driving to Dulles or BWI.

Thank you! I’ll talk to the startup that’s hiring me soon to set up a consult with you through formal channels. Appreciate the responses here!

There’s a cap on how many EB1 green cards they give out per year, and there’s also a per country cap within that cap (no country cannot cross 7% of total cap)

Unfortunately, there are a lot of EB1C greencard applicants from my country (India) as the bar for L1 and EB1C is pretty low. Since all EB1 visas (EB1A, EB1B, and EB1C) are counted together towards the cap, that means that even though I required a muncher higher bar to qualify for EB1A, I have to wait in queue with all the EB1C applicants who applied before me.

Based on my estimate, it’ll be more than 6 years (or even more) for EB1A priority date to come to late 2024.

Thank you so much for the response! Could you share a bit more on what the preparation looks like? It’s something I’ve seen vaguely mentioned in forums but no one shared the specifics of the prep. Thanks!

Hey Peter! Thanks for doing this!

I’m trying to understand the dual intent nature of O1A. I’m a PhD student with EB2-NIW and EB1A approved. Unfortunately, I am not expecting a green card in the next 6-12 years due to my country of birth.

I am eligible for O1A, but it seems like the dual-intent of O1A seems to be in grey area unlike H1B. Would I have any issues with O1A stamping in foreign embassies or during reentry given that I filed I-140 and showed immigration intent? Or is it safe to assume that O1A is also dual intent? Thanks!

Given how widespread PyPI usage is, I'm surprised they only have one full time security staff. I mean I guess it makes sense, usage doesn't always mean they get more donations/money, but damn.

As someone with admittedly no formal CS education, I've been using conda for all of my grad school and never managed to break it.

I create a virtual environment for every project. I install almost all packages with pip, except for any binaries or CUDA related things from conda. I always exported the conda yaml file and managed to reproduce the code/environment including the Python version. I've seen a lot of posts over time praising poetry and other tools and complaining about conda but I could never relate to any of them.

Am i doing something wrong? Or something right?

Someone I know got it recently. If you're getting with the API and all the sensors it was close to 60k including customs and all of that.

My theory as to why all the bigtech companies are investing so much money in video generation models is simple: they are trying to eliminate the threat of influencers/content creators to their ad revenue.

Think about it, almost everyone I know rarely clicks on ads or buys from ads anymore. On the other hand, a lot of people including myself look into buying something advertised implicitly or explicitly by content creators we follow. Say a router recommended by LinusTechTips. A lot of brands started moving their as spending to influencers too.

Google doesn't have a lot of control on these influencers. But if they can get good video generations models, they can control this ad space too without having human in the loop.

There isn't a significant upside either.

At the end of the day, buring cable throughout it's way is very expensive while damaging it at one spot, even if it's buried, is not that hard or expensive.

What makes you say TPU has zero chance against growing NVIDIA?

If anything, now is the best time for TPU to grow and I'd say investing in TPU gave Google an edge. There is no other large scale LLM that was trained on anything but NVIDIA GPUs. Gemini is the only exception. Every big company is scrambling to make their own hardware in the AI era while Google already has it.

Everyone I know who worked with TPUs loves how well they scale. Sure Jax has a learning curve but it's not a problem, especially given the performance advantages it gives.

I mean Jeff Dean is probably more ML researcher than probably 90% of the ML researchers out there. Sure, he may not be working on state of the art stuff himself; but he's too up the chain to do that.

Other countries, maybe. But China is never gonna conquer or do reasonable business in India for geopolitical reasons.

Oh yeah of course, I personally have a physical screen that shows live feed from my local transit agency's GTFS real time feed. I was just pointing out the app for those who don't have the time/interest to build something physical.

This sounds like a fun project, but there are existing apps whose default page is "when is the next bus/train coming up on stops near me"

Transit App (https://transitapp.com/) is one of them and I freaking love their interface overall. This app's default view shows you the next bus (in either directions) at the 3-4 transit stops closest to your current location. And you can customize/add favs too. It's a beautiful app, also allows for multi/mixed-modal route planning (part walking, part bike, part bus).