Company description: We do construction compliance AI. We're in series B, about $40m raised. Growing a lot! Feel free to reply or DM me if you want some more info.
Tech Stack: Typescript full stack except for Python for ML, NestJS backend, NextJS frontend, React Native mobile, IaC in Pulumi using TS.
Open Jobs:
- Senior/Staff/Principal Frontend Engineer (fullstack Typescript nice to have)
- Senior/Staff/Principal Node Engineer (fullstack Typescript nice to have)
- Senior/Staff/Principal QA Engineer
Location: Austin or Texas preferred, but open to remote for the right candidate in the US.
I used to work in ASR. Due to the nature of current multimodal architectures, it is unlikely we'll ever see accurate timestamps over a longer horizon. You're better off using encoder-decoder ASR architectures, then using traditional diarization using embedding clustering, then using a multimodal model to refine it, then use a forced alignment technique (maybe even something pre-NN) to get proper timestamps and reconciling it at the end.
These things are getting really good at just regular transcription (as long as you don't care about verbatimicity), but every additional dimension you add (timestamps, speaker assignment, etc) will make the others worse. These work much better as independent processes that then get reconciled and refined by a multimodal LLM.
Two PMs left a company with 125,665 employees. Yes, they are very important, but this happens every day at companies all around the world. Could be burnout, they are rich, got better offers, anything really.
People are so desperate for Tesla to fail they will latch onto anything.
What is the specific criticism? You believe it should not be reported on? Or you simply don't like the tone (in which case, why are you omitting that)?
He wasn't the "head of cybertruck" - if you look at his LinkedIn, he was a lead program manager. Still an important cog in the machine, but he'd been there for 8 years from an intern.
Model Y guy makes it sound like he was more of a key person, though.
> I wish .NET was more popular among startups, if only C# could get rid of the "enterpisey" stigma.
Too hard to ignore the benefits of cross-stack gains in Typescript/Python. The C# native phone, Blazor, etc just isn't quite there yet. Tried it at the last company, and full stack TS was just so much easier to do.
The reality is that the vast majority of startups don't make it. The #1 thing startups should be focusing on is hiring the right people and product velocity. TS just makes that easier in my experience.
Most of your electronic devices work with embedded software. Production lines, transport gates, cranes, computer hardware, ships, planes, rockets, cars, e-bikes, smart lights...
There is also scientific programming, that feeds research and analysis. Weather reports? Statistics, etc.
And there is gaming.
Devops, infrastructure? Databases? Tools for artists? Most of those aren't web. And yes I've heard of Figma.
There are probably tens of categories I'm missing.
Web is still bigger probably, but I have a problem with the saying "practically all other development is web".
I didn’t deduce it from the name. I deduced it from over a decade experience working primarily with .Net.
Since I can’t presume the reader has equivalent experience, especially in HN, pointing to the name, which should be a good signifier of what something does (to be fair, MS really, really sucks at naming), is a good shorthand.
The real reason .Net isn’t good for embedded devices is because MS didn’t develop it for embedded devices. They’ve only added low level memory management in the past few years.
Until recently you couldn’t even create a fully statically compiled executable.
Define “practically all”. I would accept “clear majority”.
But practically all? Nah. I mean the hot new areas for funding right now are AI and robotics neither of which are web!
I’m coming up on 20 years professional experience. Exactly none of it has been mobile or web! The programming field is so much bigger than HN likes to pretend.
>I mean the hot new areas for funding right now are AI and robotics
Most developers are not in such startups. There is a lot of boring software out there which is a website. Even for AI, the first company that comes to mind OpenAI is known for ChatGPT, a web product. Most of the AI companies are building web products.
It's web in a (limited) sense that there's probably a web frontend somewhere, but this "somewhere" is usually pretty far away from where most of the code is developed.
Most of the backend logic is not related to serving data for the browsers, it's doing actual backend stuff - communicating to databases, APIs, etc.
Is Google search backend a web app? I think it's really stretching the term.
Is it though? Backends can be any language and there's a lot more variety there -- TS+node, Go, Python, Java. It's just .NET that's largely ignored for no real technical basis.
It really depends where you are. In the UK half the places seem to use .NET in some form or another.
I am pretty language agnostic and I am reasonably competent programming in C# (I worked with C# and VB.NET for about 15 years), Go, Python, TypeScript and C++ these days.
The issue with a lot of places that do C#/.NET stuff is that they will typically ignore new tech until it is officially blessed by Microsoft. You can have a piece of tech that everyone is using and works really well and it will be ignored if it isn't blessed by Microsoft.
The other issue with .NET is all the Microsoft gumpf that tends to come with it even with the newer versions of .NET.
I am also in the weird place of being a Linux user. I've had job interviews that wanted to do live coding exercise/take home code exercise and they expect you to do everything in Visual Studio with SQL Server.
They don't even know Rider exists a lot of the time. It is also quite different visually compared to Visual Studio code.
A lot of places have never used Linux at all and if they have they have it would be WSL or some RHEL box. So if you are screen sharing Gnome and with a totally different IDE and Terminal the person assessing you might not actually understand what you are doing.
> At least you can run SQL Server easily on Linux using docker.
1) They normally want you to use something like SQL Server Compact or SQL Server Express and a specific version. TBH I just don't bother anymore with these interviews because it takes like a couple of hours to get all this stuff working on Windows.
2) SQL Server Projects can only be used on Windows with Visual Studio. Some places do a lot of stuff "old school" and they want you to use that.
> It's just .NET that's largely ignored for no real technical basis.
As someone who has been developing primarily on .Net for the past decade this is absolute bullshit.
1. It’s only very recently that .Net became open source. Until then you would frequently hit issues where the only option was to rely on the few support calls you got with MS engineers with your $1000+ Visual Studio subscription to move forward. And believe me, this isn’t a pleasant way of debugging.
2. It’s only recently that .Net became cross platform. Until recently .Net meant you had to pay far more money for windows servers, get far less performance, and open your application to way more security issues. And when things broke they broke in highly inscrutable ways.
3. It’s still not a great platform. If you’re deploying on Windows, there are still things you will want to do that will require windows registry changes.
4. It’s only recently that the transition to an open source/cross platform framework has stabilized. Until now you had to deal with MS alphabet and naming goop, an absolutely muddy roadmap, and if you ever got thrown into a project you’d end up finding yourself in a mess of varying conventions, project types, incompatibilities, etc.
5. You know all those performance improvements they’re delivering with every release? There’s a reason for that. Until recently performance was so bad. Kestrel alone provider at least an order of magnitude of improvement.
6. Thank the lord for Jetbrains but other than them, to do proper .Net development you need to use Visual Studio. And Visual Studio is not a pleasant IDE or development environment at all.
There were a lot of technical reasons to not adopt .Net. Even today there’s the problem of MS losing interest or making the wrong choices and there being basically no alternative, because unlike even Java, the .NET ecosystem is completely dependent on what MS does.
By recently you mean a decade ago yeah? I mean it’s fair that it was only a half-decade (.NET 5) when it was genuinely complete enough, but lots of stuff was in good shape when it was called .NET Core.
It sounds like you’re projecting the problems of an existing .NET shop onto the shape of a startup without all that baggage. I can assure you, having worked with many customers running new business on newer .NET, it hasn’t been a legit technical concern since about .NET Core 3.
A decade ago is when they started the transition. It’s been painful.
If you’re a new shop that is making decisions without looking into how the company that pretty much runs the platform you’re basing your future on has acted in the past decade (we’ll ignore how they’ve acted beyond that because then it’s a no brainer) then you’re doing yourself a disservice.
I see that you’ve narrowed the goal posts to just technical concerns, which is fair, but isn’t sufficient to make a decision about what technology to choose.
Especially in a field where you have a similar alternative in Java where the sponsoring company doesn’t have half as much control, as well as several fully open source alternatives.
While suffering the performance loss of V8 versus CLR, JVM or any compiled language.
One of the reasons I am back to writing more C++ code is C++ addons for node.js, as several SaaS products now only care about Next.js as extension SDK.
I don't think it's about dreaming to be Google. K8s is pretty easy to set up now with a hosted cloud platform if you start with it, and helm takes care of pretty much all your infra needs. Migrating to K8s is what's awful. From there, the docs have most everything you need to know and there's an abundance of helpful information online that covers most problems you'll run into.
I would love this to be true, but it isn't. I've done generating types for the frontend multiple times, sometimes from C# (around 2016, using typelite), Java (openapi template generator) and most recently straight from OpenAPI spec files (.yaml) using Orval.
It always has been a shitshow. It works well for the 90% cases, but in the 10% edge cases, things break. It becomes impossible to fix generation issues, you will often resort in working around issues in your backend/openapi code. Sometimes you report bugs upstream and hope it gets fixed. In the current project we are stuck on a ~2year old Orval version (a typescript generator from openapi) because some features broke or were removed in the latest version, and the entire monorepo (15+ LoB apps) wouldn't compile and would require major changes. This simply because a never version of the generator was broken/removed features previously present.
No, that's not true. If you share code like this then you can do things like put the same validation code in the frontend and the backend: frontend to give a nice user experience, and backend to protect the endpoint.
OpenAPI does support patterns for fields and nullables/non-nullables - that already gets you very far regarding validation. A decently sophisticated generator (which don't exist IMHO) would generate the validation code for your respective language.
Still one lang on both ends is nice: there are some bits of code you want to run on both ends (like templating for SSR/SEO/caching; but also using them in the browser).
Not really? Having come deom a TS + Go startup it’s pretty trivial to wire up domain objects across each language and define a clean API boundary with some enforcement at build time. And Go was a far better choice for the backend than TS for some lower-level memory considerations.
With TS you don't have to "define a clean API boundary with some enforcement at build time". You can use the exact same types you're using to enforce endpoint schemas everywhere.
You don't need a middle man like OpenAPI (which I've used and it's a mediocre solution).
Company description: We do construction compliance AI. We're in series B, about $40m raised. Growing a lot! Feel free to reply or DM me if you want some more info.
Tech Stack: Typescript full stack except for Python for ML, NestJS backend, NextJS frontend, React Native mobile, IaC in Pulumi using TS.
Open Jobs:
- Senior Frontend Engineer (fullstack Typescript nice to have)
- Staff Analytics Engineer
- Senior Product Designer
Location: Austin or Texas preferred, but open to remote for the right candidate in the US.
This is kind of a wild story. Sam Altman is openly flexing that he was able to skirt the rules and regulations by threatening economic damage to California. It's not even subtle anymore.
This reminds me of when the former CEO of Hyundai, Chung Mong-koo, went to prison for embezzlement. In just 3 years he was pardoned because the President of South Korea basically said, "we need you for the economy."
We're not even pretending that the government is in control anymore. It's just full on anarcho-capitalism on display.
whoa there. corporations have the right to move to the best location for them. California does not have the eternal right to OpenAI's taxes and employee base. think about what you're implicitly assuming here. if one company simply leaving causes concerning "damage" then perhaps it is the government that is the problem, not the corporation driving economic growth.
Corporations are applications of government power and have no natural rights, their legal rights are at best proxies for the rights of natural persons with some interest in them, and at worst a fiction which serves to limit the rights of natural persons.
Also, having a right to do something does not contradict a description of you leveraging power by using a threat of doing it, in the first place.
> Corporations are applications of government power and have no natural rights
Sure. Whatever. The people who own and work at OpenAI have no obligation to remain in California.
(I’d also argue that global norms are currently walking back from the notion of natural rights pretty much everywhere except for in some parts of Europe. The concept doesn’t work without an appeal to divinity.)
The legal entity is what owns the IP, contracts, obligations, property, etc… so even if 100% of the current employees and owners moved, their collective worth and prospects would be significantly reduced if the legal entity doesn’t also come along.
There is a book written about this topic. SA is the classic manipulator. He talks to each person whatever they want to hear, irrespective of what he really believes. Many of the things he says about AI and OpenAI are clear BS, but he'll sell that to anyone with a different flavor.
(0) Be exceptionally intelligent and capable of applying that intelligence to people, not just code or math — necessary for everything that follows.
(1) Keep attention diversified as long as possible until the winning path becomes obvious.
(2) Focus on fringe bets, but pursue many simultaneously until one clearly dominates (see (1)).
(3) Extreme social manipulation — people-pleasing, control- and power-seeking, selective transparency, skillful large-scale dishonesty, and a willingness to hurt or betray when it serves (1) and (2) and the relational cost is acceptable.
(2) brought him into the startup ecosystem and the first YC batch in the first place (he had to start somewhere); combined with (3) he made an early fortune from a failed startup. (3) also ingratiated him with PG and others in those years. (1)+(2) ensured he always had exposure to every plausible frontier of the industry; when he was effectively fired from YC for (1)+(3), (2) made the OpenAI pivot the next obvious move — and a better one. (3) almost cost him his career a second time when the board fired him from OpenAI temporarily, but he survived because (3) also ensured he had enough to offer everyone else that he leveraged his way back in.
I can only speculate so take it with a spoon of salt... But people who are really good at sussing out other people's wants can find ways to make more people happy. It's like a maze of interests and I guess he's just really motivated and good at navigating and aligning them (at least among the rich and powerful).
He also seems to understand something about power and perception, in that he takes calculated risks that seem to keep working out.
So in other words, he seems to be an extraordinarily skillful politician (in both the general and the Patrick Lencioni sense).
It's too early to say if his risks "keep working out". Restructuring is not a risk. His, and others', original decision to make the company a non-profit was also not a calculated risk in this sense.
When he was fired from OpenAI, his use of employee manipulation to regain his position is not a risk; it is the only option he had. It was his bond maturing, of carefully cultivated loyalty he had accrued over years. Gaining that loyalty was not really a risk. It was smart politics.
One risk he took is: signing away such a large portion of the company to Microsoft. I'm not sure whether that is working out.
Another risk he took is: neglecting and sidelining the "safety" portion of his organization. This caused a talent exodus and led to the formation of many competitors. I'm not sure whether that is working out either.
This is the guy who had everyone here rooting for him when the OpenAI board tried to kick him out. Which was almost certainly the result of a very successful PR campaign carried out by him.
He's not really comparable with Jobs. This guy is a politician and Jobs was a product guy.
Lol what? You need to do more research fella. Did it not ever occur to you that Jobs got the recording industry to dance to his tune to offer music for under a dollar? He did the same with the iPhone - smashing the control cellular network providers had over handset producers.
"This is the guy who had everyone here rooting for him when the OpenAI board tried to kick him out."
Finally... oh this is just funny. Who rooted for S Jobs when he was kicked out of Apple, and cheered for him on his return?
It's a bit much to say he goes into situations with nothing - he, along with others created OpenAI. It's a decent sized company now and they have sway with governments. All decent sized companies do as employers. This kind of stuff happens all the time and it just doesn't get reported as much.
It's a great question. Here are two Paul Graham (PG) quotes on Sam Altman (Sama) from 2008 and 2009.
Note, PG is the founder of YC, Sam's former boss, and the one who removed Sam from the position of President of YC after first appointing Sam to succeed him as President of YC. (Sama was more focused on OpenAI than on YC at the time, which doesn't work when you're supposed to be leading YC.)
Sam Altman has it. You could parachute him into an island full of cannibals and come back in 5 years and he'd be the king. If you're Sam Altman, you don't have to be profitable to convey to investors that you'll succeed with or without them. (He wasn't, and he did.) Not everyone has Sam's deal-making ability. I myself don't. But if you don't, you can let the numbers speak for you.
5. Sam Altman
I was told I shouldn't mention founders of YC-funded companies in this list. But Sam Altman can't be stopped by such flimsy rules. If he wants to be on this list, he's going to be.
...
What I learned from meeting Sama is that the doctrine of the elect applies to startups. It applies way less than most people think: startup investing does not consist of trying to pick winners the way you might in a horse race. But there are a few people with such force of will that they're going to get whatever they want.
from OpenAI to Helion energy (Fusion), to Retro Biosciences (longevity), Neuralink (brain computer interface), to Reddit
Sama really wants to "build the future," and when some of those investments "hit", like OpenAI did - basically become the first new company with a clear path to a $1T valuation since Facebook or TikTok), you gain immense credibility for "betting the future will happen and getting your organization there first."
If YC's motto is "build something people want," and OpenAI is now serving 800M active users while delivering incredible revenue growth (and investors want to see both). Sama gains power by giving investors what they want, by giving users what they want, and basically authoring an entire new type of software company and a new part of the economy.
A thing to note here is that, being a YC partner and top angel investor from 2011 to 2020, you can argue that Sam himself is "the most successful YC graduate." He saw thousands of companies go through YC. He saw hundreds of 'hard tech companies' go through YC. And in that decade, he could only have learned an immense amount about how VCs/successful CEOs think and make decisions. Certainly, we see the learnings of those experiences in what he's been able to pull off since.
The government believed him that OpenAI will stay in CA the same way that that government, OpenAI founders and others believed that OpenAI will be non-profit.
I think here is the answer to another commenter's question about success - it looks like great success comes to one who is able to recognize in time when old obligations become "obsolete" and thus drop them well before those obligations start to block or heavily tax the way to further success.
didn't Sam does this before with another company? he was able to restructure it and come out on top... searching for article... was it Helion? now that i think harder, didnt he comment about it on HN??
That seems like expected behavior from someone whose barely veiled pitch is "the first one to AGI will IPO - Install Planetary Overlord.[1] Give me money and be part of the future ruling elite!" (Whether his actual plan is based on what he's saying in public, or to take the money and run, permanently install his company as in inescapable rent-extracting middle-man for modern life[2], or something else, I don't know.)
1. Credit to Charles Stross, The Jennifer Morgue
2. Thneed-style capitalism
This is very cynical, and almost certainly deliberately false take.
Many situations are fundamentally uncertain with respect to laws/rules in place. The idea that they "skirted rules and regulations" is wrong. Did they push in an attempt to get a decision made which would favor them? Sure. But there was a decision to be made.
And, the people are supposed to be in control anyway, not the government.
This is Sam Altman we're talking about, a man with a long track record of conning people. I don't think any benefit of the doubt is warranted here - I'm going to assume guilty unless and until clearly proven innocent because you only get to screw people over so many times.
It doesn't matter who it is, it's naive about how the legal system works and takes the worst possible view of all the actors involved.
This wasn't some straightforward matter. Many things are not. OpenAI are not obliged to passively sit and wait while various other parties try to influence the outcome (and there were many, and they weren't nobodies).
I'm skeptical of v1 of this technology, but I could imagine a mature version of this technology could be great.
And $500/mo for essentially an always-available housekeeper seems very reasonable.
Where I live, having a housekeeper come for a few hours once a week costs about $100 a week, or $400/mo. Having a robot that could potentially always be there to:
* Tidy up.
* Clean
* Do laundry
* Help with other stuff
Seems well worth $500/mo. I don't expect that V1 of this technology will be able to effectively do all that stuff, but I'm hopeful that v2 or v5 might be able to.
On a related note, "folding laundry" seems to be a really hard challenge for machine learning to solve. Solutions like "Foldimate" kind of work if you individually hand it every piece in the right way - but nothing seems to be cable of having a human dump a bin of washed clothes in and spitting out nicely folded laundry. And everything so far that's promised to do that seems to be vaporware.
Slower is only a problem if you're waiting on the machine. I recently purchased one of those "all in one" heatpump washer and dryers. It is indeed on a per wash basis slower than my old separate washer and dryer. But over the course of a week and multiple loads, the total time spent is about the same or possibly even less.
Sure, my old washer could wash a load in say an hour and the dryer could dry that load in 2 hours. So 3 hours per load. Except that was only true for the first load. The second load has to wait for the dryer to be done with the first load, so it actually takes 2 hours to "wash" and then 2 hours to dry, so 4 hours total. And that assumes that I'm home or available at just the right moment to swap the loads. And forget running a load overnight. I mean I can, but why would I want to leave a sopping wet mass of clothes sitting waiting to be thrown into the dryer. The new one takes anywhere from 4-6 hours for a cycle to run. Seems like a terrible trade off, except I can start a load at 11 at night, and have a cleaned and dried load in the morning. I can throw a load in before I leave for work, and it will be cleaned and dried when I get home. It doesn't matter than it took an extra 3 hours because I wasn't there waiting on it, and I didn't have to swap the loads.
A side and unexpected benefit of this machine too is that it's actually faster at drying loads of bedding. The big problem with a classic tumble dryer and bedding is that it spins in one direction constantly. Early on when the bedding is all wet and heavy it starts rolling into a ball, and no matter how good your dryer's sensors are, you will almost inevitably open that dryer to a mass of hot on the outside bedding and damp on the inside. You'll unravel the mess, and throw it back in for another round or two. Because the drum unit for the all in one is the same as the washer unit, it spins in both directions while drying, just like the washing machine does. As a result, bedding never gets wrapped and balled up during the drying phase and the bedding comes out dry first time every time.
I'm skeptical too, but the fact that it works slower isn't too much of a problem if it doesn't require human attention and finishes before one is back home. It's just like how the Roomba can take as much time as it needs to to vacuum the living room when I'm gone for the day, as long as it's done by the time I get back.
I have the Xbox Ally, Steam Deck, and original Ally.
The original Ally software launch was a disaster. Unbelievable amount of bugs and overall terrible user experience. After 6+ months of updates it was decent.
I figured, hey, maybe they figured it out in advance this time? So I pre-ordered an Xbox Ally.
It is a complete disaster in terms of software. It took 90 minutes to setup and download initial updates on a Google Fiber connection. Things break constantly.
The other day, I got a new error, "Something went wrong and your PIN isn't available." When I try to click anything, it just goes black. After 6 or 7 restarts, it randomly glitches out and takes me right to desktop without any PIN.
It is just constant bullshit like this. The entire experience breaks over, and over, and over. I hate it so much. Back to Steam Deck.
I don't mean to be rude, but why would you give them even more money after screwing it up so bad the first time? You're just rewarding bad behavior at this point.
OP explained that. The original was stable after 6 months of patches. The Xbox Ally should in theory have that stability baked in. Everyone deserves a second chance. Not so much a third.
Company description: We do construction compliance AI. We're in series B, about $40m raised. Growing a lot! Feel free to reply or DM me if you want some more info.
Tech Stack: Typescript full stack except for Python for ML, NestJS backend, NextJS frontend, React Native mobile, IaC in Pulumi using TS.
Open Jobs:
- Senior/Staff/Principal Frontend Engineer (fullstack Typescript nice to have)
- Senior/Staff/Principal Node Engineer (fullstack Typescript nice to have)
- Senior/Staff/Principal QA Engineer
Location: Austin or Texas preferred, but open to remote for the right candidate in the US.
Jobs page: https://www.documentcrunch.com/careers#open-roles
reply