Alright, here we are.
Uncategorized
27
Posts
15
Posters
6
Views
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
@stefano coming from a different approach, many of the llm models know bsd internals and approaches just fine. i launched an agent (we fully control the gpu cluster at the dc) at a jail cluster staging candidate and it found the vnet misconfiguration in less than 30s.
i'm also seeing more and more companies say "we'll use ai" but the funny part is the clients i'm starting to lose to ai/llm come back with problems they can't solve since they don't have technical staff. as a result i reset my rate schedule with them at a higher cost. two have still asked about llm use, so i've shifted to building them private local-only (no calls) ai clusters for automating a lot of their tasks (think lead-gen, marketing analysis, financial tedious tasks)
i think it can be used in a constructive way, but companies see "ai/llm" as this magic wand they can wave which most shoot themselves in the foot.
i realize it's not a pivot for everyone, stefano but so far it's worked for me.
-
@stefano ok, first things first, how someone became "expert" in vibe coding?
-
As my old (Italian) boss used to tell me: «Attacca il ciuccio dove vuole il padrone…».
I do believe that you are forced to use the supercazzola AI somewhere, confined in some areas that doesn't bother you but that helps to keep your marketing alive and competitive.
-
undefined oblomov@sociale.network shared this topic
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
@stefano hosting vibe-coded services is the best way to expose security vulnerabilities. People use AI in all the worst ways and think it's capable of deep thought.
I'm studying the technology in university and I would never leave it in charge of any unsupervised critical task. At best, it can write a prototype that I'll rewrite later.
It's a glorified summary generator that just happens to generate convincing texts.
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
@stefano Vibe coding. Great. Exactly more of what the world needs.
They don't know what they don't know and that is by several orders of magnitude more than what they do know.
Ugh.
-
@stefano "experts in vibe coding"? Wow, is that really a thing? OK, I'll be honest, I let "AI" to create a short script today because:
A. is not critical
B. I was only interested in the result of it, not wanted to learn how to do it myselfBut anything related to any kind of work still created in cooperation of the hallucinating machine? Wow, that's brave.
@peterkotrcka @stefano they've been using the term prompt engineers. They think that their prompts is more likely to produce something good because they use common terminology and add make no mistake at the end. It's a cult.
-
I guess the experts in vibe coding costs lesser than a senior developers...
@freezr @stefano @peterkotrcka it ends up costing more because you pay for the tokens and you still need a full-time senior developer to verify the output. Amazon is actually having this exact problem with their new AI policy.
-
@stefano coming from a different approach, many of the llm models know bsd internals and approaches just fine. i launched an agent (we fully control the gpu cluster at the dc) at a jail cluster staging candidate and it found the vnet misconfiguration in less than 30s.
i'm also seeing more and more companies say "we'll use ai" but the funny part is the clients i'm starting to lose to ai/llm come back with problems they can't solve since they don't have technical staff. as a result i reset my rate schedule with them at a higher cost. two have still asked about llm use, so i've shifted to building them private local-only (no calls) ai clusters for automating a lot of their tasks (think lead-gen, marketing analysis, financial tedious tasks)
i think it can be used in a constructive way, but companies see "ai/llm" as this magic wand they can wave which most shoot themselves in the foot.
i realize it's not a pivot for everyone, stefano but so far it's worked for me.
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
@stefano "They are "experts in vibe coding", and for management that's more than enough."
That's rich
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
I guess that we are in the middle of a battle between dreams and reality.
CEOs dreams about 1 intern doing the job of 10 seniors. AI CEOs dreams everything done with AI. NVIDIA CEO dreams to make cash for ever.
However there is this startups (Taalas) that stated it can print a model a full model into a chip and make AI totally inexpensive, and this can a be a total game changer.
Because the risk here is to deliver the whole job market in the hands of handful of IT corps with bad reputation and arguable integrity.
-
Alright, here we are. The decline is accelerating.
An IT manager at a client company, someone capable I've been collaborating with for years, recently hired three new developers. However, he asked me for a Linux server instead of the usual FreeBSD because "that way the devs can move faster, AIs can't produce valid results for BSD systems".Given our relationship, I called him and told him I disagreed. Somewhat bitterly, he replied that these guys had been "imposed on him". They're polite and willing, but completely lacking any real programming principles. They are "experts in vibe coding", and for management that's more than enough.
In other words, we're not supposed to build a working and efficient server anymore, but a vibe-coding-friendly one.
My instinctive reaction was to ask him whether, when a data breach eventually happens, because sooner or later it will if the people writing the code neither write nor read code, they'll be able to tell the authorities that the data controller was an AI.
He didn't say anything else and thanked me. Maybe, and I stress maybe, management will understand that.
@stefano I would ask them what features they need.
this reminds me of one support call I had when I was working for Rackspace in 2013. The customer wanted Ubuntu instead of RHEL because "the dev told me Ubuntu has newer software and RHEL is too old for new stuff".... too bad all software versions were equal thanks to our backport repos... And RHEL had better support anyway (at least at the time, nowadays they are almost on par).
Feed RSS
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@Yaku ma vogliamo parlare anche di ilpost.it/2026/03/11/moschea-m… ?
(a scelta, della parte divertente popcorn, o di quanto sia agghiacciante che qualcuno possa dire che gli aderenti ad un partito che è al governo si fanno chiamare camerati, senza che nessuno faccia una piega)
-
Spoiler: "neanche nei peggiori negozi di Caracas si trova una roba del genere"
Il nuovo MacBook Neo ha una memoria SSD molto più lenta rispetto a quella degli Air | Hardware Upgrade
> Il MacBook Neo, come confermano i primi test, ha una memoria SSD molto lenta, almeno nel confronto con i risultati evidenziati dagli Air e in particolare dagli Air con M5
-
PRIDE Daily (@pridedailypod.bsky.social)
https://bsky.app/profile/pridedailypod.bsky.social/post/3mgutqh3a2k24
> Judges say, under European Union law, trans people are permitted to change their status - and that this supersedes the rules in any EU country, such as Bulgaria, that prevent legal gender recognition. >>> PRIDE Daily: Pod app · smart speaker · YouTube · pod.link/1712324163 2/2
-
PRIDE Daily (@pridedailypod.bsky.social)
https://bsky.app/profile/pridedailypod.bsky.social/post/3mgutqdtips24
> The official documents of transgender people must now recognise their gender identity across the European Union. It follows a ruling from the EU’s Court of Justice in the case of a transgender woman from Bulgaria, where courts refused to update her details. 1/2
-
Sandrino Russo (@sandrinorusso.bsky.social)
https://bsky.app/profile/sandrinorusso.bsky.social/post/3mguu3lvpns22
> Space Age 1970 Set da pranzo di Rudi Bonzanini per Tecnosalotto
-
@OrionBelt Mercoledì 26 novembre 2025...abbiamo aggiornamenti?
i soldi buttati nel referendum potevano essere meglio spesi per la gente
-
@mhoye So, a very good question about Devuan; How is it with dealing with proprietary chips, like wifi?
-
@anfangel.bsky.social
https://bsky.app/profile/anfangel.bsky.social/post/3mgu3es2l6k2e
