sigotirandolas

5 ms·

1.
by sigotirandolas 3mo ago
Perhaps that's true of the top 5% engineers. But the question is if the top 40% of engineers can replace the bottom 60%. The LLMs don't need to be perfect, they just need to be good enough so that the cost of fixing their code is
2.
by sigotirandolas 6mo ago
I don't look at whether the text is written by an LLM but at whether it has substance and whether the writer understands what they are doing and is respecting my time. If the text is full of punchy three word phrases or nonsense GenAI
3.
by sigotirandolas 6mo ago
Sorry, I think you're right that I misinterpreted your comment. I still had in mind OP's example (BDD, mutational testing, all that jazz). I apologize! Reading your comment, it looks like you work for a pretty nice company that ta
4.
by sigotirandolas 6mo ago
On the one side I reject that product and engineering concerns are separated: Sometimes you want to avoid a feature due to the way it will limit you in the future, even if the AI can churn it in 2 minutes today. On the other side perhaps yo
5.
by sigotirandolas 6mo ago
To be devil's advocate: Many of those tools are overpowered unless you have a very complex project that many people depend on. The AI tools will catch the most obvious issues, but will not help you with the most important aspects (e.g.
6.
by sigotirandolas 8mo ago
Luckily those I work with are smart enough that I've not seen a PR thrown away yet, but sometimes I'm approving with more "meh, it's fine I guess" than "yeah, that makes sense".
7.
by sigotirandolas 8mo ago
The most annoying thing is that even after cleaning up all the nonsense, the tests still contain all sort of fanfare and it’s essentially impossible to get the submitter to trim them because it’s death by a thousand cuts (and you better not
8.
by sigotirandolas 8mo ago
N~10^(6.5) aliens.
9.
by sigotirandolas 8mo ago
I hope this too but it's not a given, IMO. Previously people without technical chops failed quickly by being unable to deliver working code, now they can deliver mediocre code with the damage only becoming clear years later. It breaks
10.
by sigotirandolas 9mo ago
This looks like LLM slop (not only the writing, but the analysis itself). To start, it is based on a single machine check. It has little context, but if this was a common problem, I'd expect more data points. The MC happened 8 hours af
11.
by sigotirandolas 9mo ago
The security model is that applications run in a sandbox (e.g. Flatpak, snap) and only get D-Bus, Wayland, etc. access via restricted means (e.g. xdg-dbus-proxy). The "friction" is that Wayland developers don't want a sandbox
12.
by sigotirandolas 10mo ago
In the short term, definitely. In the long term, once the damage from vibecoding is better understood (for customer impact and team morale), there's an incentive to push them out, both from the leadership and the individuals side.
13.
by sigotirandolas 10mo ago
There's definitely some that hold CQRS, DDD, TDD, ... as _the_ way to design software and over-engineer around it, so I can understand some pushback. Knowing those patterns is very helpful as a way to think about design problems, as lo
14.
by sigotirandolas 10mo ago
When all leadership is asking is "what is the short term business value?", it's pointless to make that case. It's much easier to measure "yet another feature" than "fix the root causes of what makes our pr
15.
by sigotirandolas 1y ago
I assume it can be different for everyone. This post resonates with me, but my social anxiety mixes being sensitive to negative feedback and low self-esteem. So, you want to avoid both being disliked, but also being liked - because this put
16.
by sigotirandolas 1y ago
> Maybe things are this equal and fair on the senior, high-paying part of the spectrum I don't think the fundamental dynamics change by seniority, just that after some level there may simply be a smaller pool. From the interviewers
17.
by sigotirandolas 1y ago
> - How do the end user protect themselves at this point? Especially the average user? - Install as little software as possible, use websites if possible. - Keep important stuff (especially cryptocurrency) on a separate device. - If you
18.
by sigotirandolas 1y ago
Not the parent, but the default `npm install` / `yarn install` builds will ignore the lock file unless everything can be satisfied, if you want the lock file to be respected you must use `npm ci` / `yarn install --frozen-lockfile`
19.
by sigotirandolas 1y ago
> Measuring and feeling are not mutually exclusive. They are not mutually exclusive, but they compete to a degree. If someone's time is mostly spent on what can be measured, they can't spend time on "common sense" or
20.
by sigotirandolas 1y ago
For as much as the author may get roasted for stating the obvious, I've often seen this "measure everything" mindset, coming from those you'd think should know better than that. I've even seen this stupidity in myse
21.
by sigotirandolas 1y ago
I'd say the opposite, LLMs are a know-it-nothing machine to perfectly suit know-it-alls. Unlike a human, it isn't that hard to get the machine to say what you want, and then generate enough crap to 'defeat' any human cha
22.
by sigotirandolas 1y ago
My only answer to this is, the ones at the top up to the CEO must be mindful enough to realize this, smart enough to figure out a solution, and brave enough to act on it. Otherwise, it's a matter of time until the house of cards falls
23.
by sigotirandolas 1y ago
> What i got back was the typical overly verbose and articulate review from chatgpt or some other llm. I thought it was pretty funny that they thought it would work let alone be acceptable to do that. Did that end up working for you? I h
24.
by sigotirandolas 1y ago
What makes me not want to use Hyprland is that the code has all kind of "YOLO" tells, the kind of ones that make you wonder if something is going to happen some say... for example: - https://github.com/hyprwm/
25.
by sigotirandolas 1y ago
A lot of it is being able to make reasonable decisions under novel and incomplete information and being able to reflect and refine on their outcome. LLMs's huge knowledge base covers for their incapacity to reason under incomplete info
26.
by sigotirandolas 1y ago
To be devil's advocate, this is the kind of all-talk argument the parent was referring to. Once the paid option is available, people will demand it to be [cheaper / better / someone else] and still not pay. While I don't
27.
by sigotirandolas 1y ago
My thought is that we humans are bad (by computer standards) at arithmetic and memorization because those are not evolutionarily useful on their own. On the other hand general problem solving is, and so far any attempt to replicate it using
28.
by sigotirandolas 1y ago
I think the more realistic argument is that the model can generalize, but only by learning shortcuts (e.g. how to pattern match a problem to a likely answer) and simple algorithms (e.g. how to propagate carries in a multiplication). And t
29.
by sigotirandolas 1y ago
I'm not convinced by this argument. You can fit a bunch of books covering up to MSc level maths on less than 100MB. After that point, more books will mostly be redundant information so it doesn't need much more space for maths bey
30.
by sigotirandolas 1y ago
You'd be surprised. Off the top of my head: Many are conditioned to see `x` as a fixed value for an equation (as in "find x such that 4x=6") rather than something that takes different values over time. Similarly `y = 2 * x` c

More ›