Niels Abildgaard

We're going to have to make some changes around here 🏴 he/they

Software developer.

Looking for my community on here. Please tag me in if you see anyone talking about:

code things;
how to do software good;
org and company design;
building european cloud alternatives;
how to make the world a better place;
cool people doing neat things;
stuff;
other

location: Copenhagen, Denmark
mastodansk: https://norrebro.space/@nielsa

Public Key

npub14gfkpmm2k2e2p3jtuq2g4wwe8y8h8nrxclt4xvu04clm8yzmzy7qmg3gyr

NIP-05 Address

nielsa_at_mas.to@momostr.pink

Profile Code

nprofile1qqs25ymqaa4t9v4qce97q9y2h8vnjrmne3nv046nxw86u0anjpd3z0qprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntqy8hwumn8ghj7mn0wd68ytnwdams6dxn4g

Publishing to

relay.momostr.pink

nostr.now

relay.primal.net

Author Public Key

npub14gfkpmm2k2e2p3jtuq2g4wwe8y8h8nrxclt4xvu04clm8yzmzy7qmg3gyr

Show more details

Last Notes

2026-10-07 17:45:05 UTC

- reply

I'm a bit confused about part of this post, which talks about something in your latest video that also hit me as pessimistic, so I'm going to point to what exactly I saw in the hope that we can bottom-up this together.
The very last sentence you say in your video is "I think it might be time to… if you work in software, I think it might be time to try and figure out a way out of it."
As the last sentence of the video, this comes across, to me, as your conclusion, and it definitely does not sound limited in any way: it's a broad conclusion about the whole industry.
In your post you say, "I wasn’t writing about your job or your employer. Glad you’re in a good place, though. I was making an observation of what’s happening to many people in the industry, how their jobs are increasingly making them miserable, and pointing out that the sensible thing for an individual in that situation to do is to leave."
To me, those two statements are mutually exclusive. What am I missing?

2026-09-25 15:00:13 UTC

- reply

afaict it is only covered as text, which at least in Denmark's interpretation means it is basically exempt if a person looked at it before publishing: https://mas.to/@nielsa/117036858924329075

2026-09-23 17:33:26 UTC

- reply

I'm moving off Netlify too due to bandwidth pricing and concerns of independence.
What is your analysis?
I once met their CEO for a coffee while I was at uni doing web engineering research for my MSc and he seemed well intentioned then. That was after they became succesful but before they truly scaled. Recently it has seemed to me that they are regressing to the Silicon Valley SaaS mean.

2026-09-14 10:16:14 UTC

- reply

A data point on outsourcing dev jobs: I've seen several companies roll back outsourcing. And many never adopted it. Many large co's had net losses from their outsourcing adventures.
But some are still trying new approaches today. (Not sure I can mention compnues by name, but let's say... large Danish companies.)
Sure, we didn't get rid of outsourcing but it also isn't more common than not—at least in Denmark?
I'm not sure I see that most companies benefit from it—most tech companies I know the workings of seem unaffected.
Would love yo hear more about what you are seeing that I am not!

2026-07-29 14:58:37 UTC

- reply

Sam Altman: "Some years ago we were talking about this moment as if it was something wild and now it's here and that's crazy to think about"
Everyone: "This is what you meant? This looks nothing like what you were all talking about for years."
Sam: "Yes. But it's exciting because it feels like we're at an infldction piint wgere it could go either way right now, so we really need more support"
Everyone: "So it looks different from selfreplicating AI and it isn't inevitable. What other parts of the singularity are left??? Just the name???"

2026-07-29 14:50:39 UTC

Sam Altman: "Turns out the singularity was the friends we made along the way"

2026-04-05 11:20:01 UTC

- reply

Wardley had an interesting take on another way this is true, I think it was in a LinkedIn post: "AI" chatbots function as a transfer of values, normalizing what the designing tech bro class think should be normalized.
A sycophantic, encouraging, but subtly opinionated automaton, that slowly tugs at the opinions, values, principles of those who interact with it.

2026-04-03 16:07:18 UTC

- reply

Aight - well, feel free to put me on a list somewhere for when you figure it out 😁

2026-04-03 15:44:57 UTC

- reply

Where can I get physical copies? Not available in my regular Danish bookshops afaict

2026-03-30 21:05:48 UTC

So you're telling me that in the future, instead of developing software by writing thoughtful code, we will all be reviewing large PRs generated by applications optimized for making bugs hard to spot, especially when inattentive, say, from exhaustion induced by continuous reviewing of massive amounts of code?
Cool. Cool, cool, cool.

2026-03-29 07:09:14 UTC

- reply

deal

2026-03-28 11:22:24 UTC

- reply

Yeahh we don't fit those criteria (yet), but could maybe build something that fits if you can share the requirements

2026-03-28 06:18:13 UTC

- reply

I might be able to hook you up with some cheap VPSes hosted in Denmark if interested... we are working with a publicly owned data center that is trying to break into the private sector and care about digital/data soverignty. Let me know if relevant...

2026-03-16 13:19:04 UTC

- reply

Nice to see this written out longform!

2026-03-01 19:55:25 UTC

- reply

Got a link to that, so I know we're talking about the same thing?

2026-03-01 19:34:49 UTC

- reply

If you mean "LLMs/machine learning can be used well and have a significant impact" I agree.
But I have yet to see the case for generalized LLMs as drivers of problem solution. Happy to look at examples if you have any.

2026-03-01 19:30:07 UTC

- reply

Very good! I agree that that's part of it... I guess a,m different piece to the puzzle is whether or not people believe the hype around *the future* of LLMs (AGI, superintelligence, singularity, ...).
I understand the appeal, but am surprised by how LLMs uniquely get to get away with such lofty claims while showing no concrete promise for so long.
Like... any other tech that doesn't work but asks you to believe it might one day would not have lasted... years.

2026-03-01 19:11:36 UTC

- reply

Aye, I see it.
But I still can't make it make sense that otherwise competent people (with all the tools to evaluate actual value) keep falling for it, c.f. https://mas.to/@nielsa/116155324997527697
I may have all the data that explains it, I just can't wrap my head around it, I guess, you know what I mean? It just feels... too silly.

2026-03-01 18:57:42 UTC

- reply

Aha! Honestly not a bad point: branding, immediate availability... the metaphor tranalates pretty well.

2026-03-01 18:22:19 UTC

- reply

Every time I try to imagine a good constrained software-related case for generalized LLMs, I end up at essentially low-code platforns, which, to be clear, already exist.
People are out here paying 10x the cost of a low-code platform subscription to get a worse project, built less securely.

2026-03-01 18:19:41 UTC

- reply

Also honestly requesting good use cases for *generalized* LLMs. Do you know of any? What am I not seeing?
inb4 anything code-related.

2026-03-01 18:16:16 UTC

- reply

I understand how the dopamine-fueled feedback loops work and why anthropomorphization and novelty works to lure some people in...
But too many people who I thought were competent are lured in by this, and double down when I point out the flaws in their output post-"AI".
I don't understand the difference between the (too many) people who fall for it, and the (too few) who see through it, check the output, and remain unconvinced ...
Honestly requesting perspectives on this.

2026-03-01 18:12:12 UTC

- reply

I kinda have to have some understanding of this tech in order to advise clients and customers on how it works, but I literally cannot find good cases for this generalized code- or text generation LLMs.
I find the custom-trained Danish lexicon bot at chat.lex.dk quite useful. It is trained on paid academic experts' input, and only that, and refuses to answer when something is not in its sources. Very cool, but limited case. Very different from the claimed coming revolution.

2026-03-01 18:09:20 UTC

Every time I try to replicate or create useful cases for LLMs in technical areas I understand (including code/software development, corroborating various claims, summarizing text, etc) I have such a terrible experience convincing it to actually answer what I ask, without hallucinating or its clear training biases to shine through.
I have yet to see anyone's shared "useful output" (non-trivial tasks) without the same flaws.
I'm not surprised, but confused: people keep claiming it works??