Not surprising. I worked at NI for my first job out of university and developing software there was so painful
Ocerge 148 days ago [-]
Same, I left in 2014. I left after a couple of years to almost double my pay and not work on a Linux RT USB driver directly out of college for which I had no desire to be a SME, but in hindsight I think the coworkers I had there were smarter than anywhere else I've worked since (including Google). They paid absolutely nothing but seemed to have good culture, at least where I was at the time.
Rinzler89 148 days ago [-]
>in hindsight I think the coworkers I had there were smarter than anywhere else I've worked since (including Google)
Why do you think that is?
formerni 147 days ago [-]
This matches my experience too. I'm a software dev in my 40s, started career at NI out of college, multiple companies since then.
NI had a fantastic college recruiting process. They sent engineers (not HR people or recruiters) to college career fairs all across the country, often back to their own alma mater, to snag the best-and-brightest from good engineering programs. They then organized big interview days where something like 50 or 100 college applicants would be flown in to interview on Thurs/Friday and then be taken out to downtown Austin on Friday night.
Their interview process was technical, rigorous but fair, and avoided the pitfalls of pointless puzzle questions or of favoring elite credentials over abilities. And their managers, product and sales positions all started from that same pool of entry-level engineers, so even if you were interacting with someone who's title was like "Product Marketing Manager" they had a BSc in engineering, comp sci, physics, math, etc.
The end result was that you always had very intelligent and capable coworkers surrounding you every day. I left for the same reason many other posters have mentioned- by mid career, I was getting paid substantially less than I could make elsewhere.
a-dub 148 days ago [-]
ahh i see, now i understand why there's still no 6000 series usb driver for linux.
so annoying!
georgeburdell 148 days ago [-]
Any insight into why NI binaries were so large and interconnected relative to competitors such as Keysight?
CoastalCoder 148 days ago [-]
Sounds interesting. Can you elaborate?
etimberg 148 days ago [-]
Some highlights that I remember include:
* One project where the CI took something like 48 hours
* Trying to apply waterfall hardware engineering processes to software development
* Mostly hiring devs fresh from university so there was a ton of group think due to a lack of new ideas
* Low salaries so most of the devs left once they got promoted beyond a junior level
sidewndr46 148 days ago [-]
You forgot "cuts the salary of everyone each year, ensuring anyone who can update a resume finds a new job"
etimberg 148 days ago [-]
Never had that experience while I was there, but about a year after I left the entire office I was in was closed and everyone laid off
sumtechguy 148 days ago [-]
a friend of mine has worked there for 25+ years. I keep telling him work somewhere else and make more money...
josh-sematic 148 days ago [-]
I had similarly painful experiences. Also the project I was on was mostly in maintenance mode, with mostly junior engineers out of college on it who had little experience with that particular code. Not many tests. Lots of code archaeology and "I'm crossing my fingers I didn't just break some subset of the enormous configuration space of this tool"
etimberg 148 days ago [-]
The code archaeology was probably the most fun part for me but I had so much free time to do that because my projects were mostly maintenance as well
positr0n 147 days ago [-]
Fun NI archeology story.
When I was there in the early 2010s, I met a greybeard hardware engineer with a big chip mask poster on his cube wall. He told me they lost the schematics and all documentation to one of their first timing chips that was used in lots of products. They ended up calling the manufacturer and asking for the mask image back, and any question anyone had had to be reverse engineered from the mask.
jolt42 147 days ago [-]
Did they leave and go to CrowdStrike?
beryilma 147 days ago [-]
> Mostly hiring devs fresh from university so there was a ton of group think due to a lack of new ideas
This is a very interesting description of group think between less experienced developers. I've never thought of it this way, but it seems to have some truth in it.
Kirby64 148 days ago [-]
NI has a nickname in the industry:
NI = No Income. Keep in mind, they're also HQ'd in Austin, TX, which is essentially the most expensive city in Texas.
etrautmann 148 days ago [-]
Interesting, I’ve ended up spending thousands on NI hardware at my last four jobs
Kirby64 148 days ago [-]
The cost of their hardware and software is unrelated to the salaries they pay their employees.
Rinzler89 148 days ago [-]
That goes for every business. Their purpose is to increase shareholder value, not employee wages. The last part only happens due to market dynamics forcing them to, but they'd gladly pay engineers minimum wage if they could.
positr0n 147 days ago [-]
A lot of their products are relatively far into the long tail of hardware projects.
Enough demand in the market that it makes sense for a company to design and build them, but not so many units sold that all those design costs are easily amortized away.
osigurdson 148 days ago [-]
I'm a little surprised that a > 1 year article about an acquisition in a somewhat niche industry is front page Hacker News.
neltnerb 148 days ago [-]
I suppose it's fair to describe both as niche, even though basically every facility that requires high reliability controls (talking chemicals, energy, medical, etc) uses Emerson. They're enormous, but they do focus on one kind of thing.
It happens to also be what NI does (did) except NI did it in a way that was more accessible to education and hobbyists. Still expensive, but with things like educational toolkits using LabView as a base they have products that address the market for lower cost, lower reliability, but more flexible prototyping tools that I've never seen Emerson focus on at all.
But yeah, I knew about this a year ago because it's the kind of thing that matters a lot for my work. And since then I've known to not build any new prototypes that use NI software and instead move towards anything else...
LeifCarrotson 148 days ago [-]
> And since then I've known to not build any new software that uses NI software[sic] and instead move towards anything else...
Do you mean "not to build any new software that uses NI hardware", or are you specifically averse to Emerson's software dev practices even though you trust them to produce good hardware?
On that note, who is your new preferred vendor for DAQ hardware? Some of the stuff that NI allowed you to build with cRIO or PC-based multifunction DAQ hardware like their PCIe-6321 etc. was pretty unique. There's not a lot of off-the-shelf gear for on the order of $1000 that can do 100 kHz digital/analog signal acquisition.
I like Delta Tau PMAC gear for electronic servo control (though their recent acquisition by Omron seems to be having a similar impact as I expect Emerson is having at NI) and Delta Motion for hydraulics (not yet bought out by anyone, they seem to successfully transition to employee-owned after Natchwey retired last year, but time will tell)... but neither is a true multifunction DAQ system like NI.
applied_heat 148 days ago [-]
Nothing like asking a question on a forum and Natchwey answering with 3 pages of mathcad solving 7th order motion equations
TheJoeMan 148 days ago [-]
At my company we’ve had success with Advantech (Japanese) DAQs.
I always considered NI best for research labs / benchtop measurements, and this acquisition seems like a poor fit because NI was pricy but with good end user support, while Emerson likes to have very expensive/little support.
neltnerb 148 days ago [-]
Well, I'd say that NI hardware was at one point the best you could get in terms of performance and ease of use, if you had tons of money. Quick prototypability is the main advantage compared to everything else.
I'm not sure who I would use for that purpose, I'd say that we're blessed with a huge variety of things which are of quite good enough performance and ease of use, but also with the curse of needing to deal with different vendors for different kinds of uses. I wouldn't expect most companies that sell general purpose DAQs to also sell FPGA systems for signal processing, for instance. That's odd.
But if you just mean digital and analog inputs and outputs there are so many options now that are adequate, even just an eval board for a microcontroller running a LabView firmware module can do well enough (and this route lets you escape NI better later anyway). This is a kind of cool new thing, using an Arduino or whatever as if it were an NI DAQ.
So the ecosystem is much larger, and less locked into NI software (LabView anyway).
On the other end where reliability is important, I wouldn't trust many places, but I barely trust NI. Honestly I'd be looking at Emerson if I wanted that level of assurance, but I wouldn't use that for prototyping. In a prototype I have to design it to fail safely enough rather than trying to make it foolproof.
So for prototyping there's a ton of options now, and once the prototyping is done -- if it needs to be high reliability then it's probably going Emerson or something similarly expensive anyway, and if it needs to be low cost then you'll just put it on an embedded DAC and ADC with a ARM-M0 or RISC-V or Raspberry Pi or whatever you fancy. And if you want it to be more expensive but more reliable you buy some kind of OEM module you put on your PCB so that they can be more carefully integrated while using certified analog input and output modules.
I think that ends up being three widely varying use cases, and prototyping is maybe just too easy to do in other non-LabView and non-NI-hardware systems such that the high-reliability Emerson-style market is what remains. I won't miss LabView, that's for sure.
kragen 148 days ago [-]
i'm not sure electronic test equipment is really that niche here
guerrilla 148 days ago [-]
They seem pretty big, but maybe people are confusing them with to old famous National Semiconductor for some reason (sold to TI in 2011.)
bobmcnamara 148 days ago [-]
It's becoming clear it's the end of a language.
ChuckMcM 149 days ago [-]
I must have missed this when it happened last year (April apparently). Perhaps not surprisingly I associate the name "Emerson" with comically large integrated stereos (boom boxes). But apparently they are building up a test and measurement group.
quickthrowman 148 days ago [-]
Emerson Electric (in TFA) does not make consumer electronics, that’s another company called Emerson Radio Corporation.
ChuckMcM 148 days ago [-]
I am aware, and it is unfortunate for Emerson Electric in terms of branding things in the electronics space. I don't doubt they will retain the National Instruments and/or Digilent brands for that reason.
It will make me chuckle every time I see that Digilent banner "An Emerson company."
148 days ago [-]
bad_username 148 days ago [-]
I spent a minute in a state of mild shock, thinking it's Native Instruments that got acquired...
svaha1728 148 days ago [-]
They were bought by a private equity company in 2021. Would love to know more about the Berlin programming scene in the 90s. The initial minds behind Generator/Reaktor were incredibly inspiring.
RIMR 148 days ago [-]
That would be a ridiculous acquisition. The shock was justified.
sixthDot 147 days ago [-]
well, remember that Sonic Foundry was acquired by Sony.
Lammy 148 days ago [-]
Interesting to see a two-character dot-com domain.
Why do you think that is?
NI had a fantastic college recruiting process. They sent engineers (not HR people or recruiters) to college career fairs all across the country, often back to their own alma mater, to snag the best-and-brightest from good engineering programs. They then organized big interview days where something like 50 or 100 college applicants would be flown in to interview on Thurs/Friday and then be taken out to downtown Austin on Friday night.
Their interview process was technical, rigorous but fair, and avoided the pitfalls of pointless puzzle questions or of favoring elite credentials over abilities. And their managers, product and sales positions all started from that same pool of entry-level engineers, so even if you were interacting with someone who's title was like "Product Marketing Manager" they had a BSc in engineering, comp sci, physics, math, etc.
The end result was that you always had very intelligent and capable coworkers surrounding you every day. I left for the same reason many other posters have mentioned- by mid career, I was getting paid substantially less than I could make elsewhere.
so annoying!
* One project where the CI took something like 48 hours
* Trying to apply waterfall hardware engineering processes to software development
* Mostly hiring devs fresh from university so there was a ton of group think due to a lack of new ideas
* Low salaries so most of the devs left once they got promoted beyond a junior level
When I was there in the early 2010s, I met a greybeard hardware engineer with a big chip mask poster on his cube wall. He told me they lost the schematics and all documentation to one of their first timing chips that was used in lots of products. They ended up calling the manufacturer and asking for the mask image back, and any question anyone had had to be reverse engineered from the mask.
This is a very interesting description of group think between less experienced developers. I've never thought of it this way, but it seems to have some truth in it.
NI = No Income. Keep in mind, they're also HQ'd in Austin, TX, which is essentially the most expensive city in Texas.
Enough demand in the market that it makes sense for a company to design and build them, but not so many units sold that all those design costs are easily amortized away.
It happens to also be what NI does (did) except NI did it in a way that was more accessible to education and hobbyists. Still expensive, but with things like educational toolkits using LabView as a base they have products that address the market for lower cost, lower reliability, but more flexible prototyping tools that I've never seen Emerson focus on at all.
But yeah, I knew about this a year ago because it's the kind of thing that matters a lot for my work. And since then I've known to not build any new prototypes that use NI software and instead move towards anything else...
Do you mean "not to build any new software that uses NI hardware", or are you specifically averse to Emerson's software dev practices even though you trust them to produce good hardware?
On that note, who is your new preferred vendor for DAQ hardware? Some of the stuff that NI allowed you to build with cRIO or PC-based multifunction DAQ hardware like their PCIe-6321 etc. was pretty unique. There's not a lot of off-the-shelf gear for on the order of $1000 that can do 100 kHz digital/analog signal acquisition.
I like Delta Tau PMAC gear for electronic servo control (though their recent acquisition by Omron seems to be having a similar impact as I expect Emerson is having at NI) and Delta Motion for hydraulics (not yet bought out by anyone, they seem to successfully transition to employee-owned after Natchwey retired last year, but time will tell)... but neither is a true multifunction DAQ system like NI.
I always considered NI best for research labs / benchtop measurements, and this acquisition seems like a poor fit because NI was pricy but with good end user support, while Emerson likes to have very expensive/little support.
I'm not sure who I would use for that purpose, I'd say that we're blessed with a huge variety of things which are of quite good enough performance and ease of use, but also with the curse of needing to deal with different vendors for different kinds of uses. I wouldn't expect most companies that sell general purpose DAQs to also sell FPGA systems for signal processing, for instance. That's odd.
But if you just mean digital and analog inputs and outputs there are so many options now that are adequate, even just an eval board for a microcontroller running a LabView firmware module can do well enough (and this route lets you escape NI better later anyway). This is a kind of cool new thing, using an Arduino or whatever as if it were an NI DAQ.
So the ecosystem is much larger, and less locked into NI software (LabView anyway).
On the other end where reliability is important, I wouldn't trust many places, but I barely trust NI. Honestly I'd be looking at Emerson if I wanted that level of assurance, but I wouldn't use that for prototyping. In a prototype I have to design it to fail safely enough rather than trying to make it foolproof.
So for prototyping there's a ton of options now, and once the prototyping is done -- if it needs to be high reliability then it's probably going Emerson or something similarly expensive anyway, and if it needs to be low cost then you'll just put it on an embedded DAC and ADC with a ARM-M0 or RISC-V or Raspberry Pi or whatever you fancy. And if you want it to be more expensive but more reliable you buy some kind of OEM module you put on your PCB so that they can be more carefully integrated while using certified analog input and output modules.
I think that ends up being three widely varying use cases, and prototyping is maybe just too easy to do in other non-LabView and non-NI-hardware systems such that the high-reliability Emerson-style market is what remains. I won't miss LabView, that's for sure.
It will make me chuckle every time I see that Digilent banner "An Emerson company."
See ti.com - Texas Instruments
Indeed. There's also this one letter dot-com with 500 million MAU: x.com.
I wonder what took so long to rekt themselves into an acquisition after Dr. T retired.