• 5 Posts
  • 81 Comments
Joined 27 days ago
cake
Cake day: May 14th, 2026

help-circle

  • Scoping first, code at bottom. Kewl?

    Is this a “learn it by coding it” project or is it a “I want this thing to exist, no one has done it, but my code skills aren’t quite there” project?

    If the latter, would you consider iterating via llm (as you mention n8n, so I figure you’re in that space anyway) or is this a purely a learn by doing thing?

    Come to think of it, there is actually a third option here. You could get the LLM to teach you how to code it by writing some pseudo code and asking it for pointers / starting steps. Claude web is pretty good for that sort of thing, I think. You can get it to tailor its lessons to what you need without the tedium of starting at “Hello World”.

    You seem like the sort that could keep that interaction honest and not let it just do everything.

    PS: I read about your setup - sounds brilliant. Go you good thing.

    PPS: n8n has a Code node (JavaScript), and parsing that weather JSON into a formatted string is probably like 15 lines of code. Something like -

    const data = $input.first().json;

    return [{ json: { temperature: data.list[0].main.temp } }];

    add a Code node after your HTTP request in n8n. Get one single value out first. If you see a temperature number in the output, you win.

    Pulling Python in just to parse JSON is probably adding a tool you don’t need for this.

    JavaScript may be ass but it’s literally there, so it’s omnipresent ass. :)

    Once you’ve got the JSON parsed, turn it into one small HTML weather card.

    const data = $input.first().json; const item = data.list[0];

    const html = <div> <h3>Weather</h3> <p>Current temperature: ${item.main.temp} °C</p> <p>Feels like: ${item.main.feels_like} °C</p> <p>Humidity: ${item.main.humidity}%</p> <p>Condition: ${item.weather[0].description}</p> </div>;

    return [{ json: { html } }];

    (Sorry about the formatting ; Lemmy formatting is weird AF)

    Anyway, as one journeyman to another, that’s where I’d start poking. ICBW.


  • SuspiciousCarrot78@aussie.zonetoSelfhostedKittygram v1.1 has released
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    13 hours ago

    Let’s tag it as “provisional” then. As in, once you have my provisional trust, accrued over time, I’ll probably stop auditing every single line. I’ll still look tho.

    But the long and short of it is this - XZ utils backdoor actually makes case for trusting clankers more than human collaborators. Clankers are incompetent… they usually aren’t Machiavellian.

    I’ve heard it said that an LLM is like a Labrador retriever when it comes to coding. Overly excited, pulls ahead, does some really goofy shit and sometimes chews up your couch (hello Qwen 27B)…but it is trainable.

    Human devs are like cats…which is oddly on brand for this project :)

    I’d sooner trust a clanker I had prompted with my house style ticket and narrowly sandboxed than a rando online. Of course, the difference is, a rando may eventually earn trust…a clanker doesn’t - but it doesn’t need to if narrowly scoped.

    EDIT: here’s a template I use / created for Qwen / Codex. It’s…opinionated and bears scars of prior over eager Labradors. This is usually step 1 I fill out. My fingers are going to shit with O/A , so am trying to minimise scut work.


    TICKET-Px-SHORT-DESCRIPTIVE-NAME

    Status: PROPOSED Timestamp: DD-MM-YY-HH-MM Priority: P0 | P1 | P2 | P3

    Purpose

    One paragraph:

    • what changes
    • what does not
    • whether this is proposal / proof / implementation

    Why this exists

    Describe:

    • concrete failure mode
    • why current behaviour is wrong
    • why this is architectural not cosmetic
    • why local patches are rejected

    Include: We do not want … We do want …

    Proof requirements before implementation

    Hard gate.

    Before implementation exists, prove:

    • seam exists
    • ownership is correct
    • contract can be enforced
    • no god-object expansion
    • no hidden coupling

    If proof fails: stop and escalate. Do not patch.

    Gates

    • Step 0 GO/NO GO
    • Step 1 GO/NO GO
    • Step 2 GO/NO GO
    • Step N GO/NO GO

    Each gate:

    • exact thing being proven
    • explicit stop condition

    Test Plan

    Mix of:

    • unit fixtures
    • regression replay
    • smoke coverage
    • edge cases
    • negative cases

    Prefer: prove behaviour changed, not just coverage increased.

    Definition of Success / PASS

    Minimum acceptable state.

    Must describe:

    • observable outcome
    • old failure closed
    • contract enforced
    • ownership preserved

    Definition of Success / EXCELLENT

    Stretch target.

    Usually:

    • generalises across adjacent lanes
    • demonstrates reuse
    • proves contract not logging theatre

    Assumptions

    State assumptions explicitly.

    Examples:

    • baseline already proven
    • implementation surface bounded
    • no broad whitelist/regex fix

    Proposed shape

    Describe:

    • modules
    • packets/cards/contracts
    • ownership boundaries
    • interfaces

    Prefer: small typed objects.

    Thin leaf intent

    If adding logic:

    prefer:

    • thin leaf
    • compact return object
    • narrow ownership

    Avoid:

    • diagnostic fluff
    • local maxima

    Policy versus signal

    Policy: config

    Signal: code

    Config controls behaviour. Signal detects reality.

    Scope

    Explicitly include:

    • what this ticket covers

    Non-goals

    Explicitly exclude:

    • unrelated cleanup
    • opportunistic refactors
    • god-object growth
    • broad routing changes

    Acceptance criteria

    Numbered list.

    Must be testable.

    Definition of done

    Agreement on:

    • ownership
    • interfaces
    • config surface
    • enforcement point

    Only then may implementation tickets follow.




  • Ironically, with a bit of effort, it’s quite possible to get an LLM to mimic your own house style and replicate it quite well.

    You can use it as an editor or hostile reviewer or just bounce ideas off it, as long as you own the process.

    Language transformation is literally what they exist for.

    But the fact that the article failed to do this and produced generic slop is the actual insult. Not the fact that they used AI.

    Like, we get it - portfolio slop / LinkedIn brain / look at me, look at me.

    I don’t mind you using AI, but could you at least steer it a bit and do more than minimal effort? Or should I get my ai to talk to your ai?


  • True…but the arguably better / more defensive stance is “accept no PR unless the user explains wtf it does and/or I personally trust them”.

    Iow, stop accepting PRs from randos - clanker or meatbag - full stop. The lowest cognitive load is “none”.

    I don’t know you / we can’t have a convo why you sent me this? Into the bin.

    (In my humble opinion, for a small or new project, that’s a cleaner footing anyway)

    The claude.md file is cute, but I don’t think a claude would actually be tripped up by that.

    It’s not such a high bar to pass to be honest with you. You’d probably need something more subtle, at which point you’re just shooting yourself in the foot.

    The meow thing is more like a philosophical line in the sand than anything else and I respect it.

    But given the way that Codeberg actually blocks crawlers and agents (and how Claude works), it probably doesn’t really do what we think it does.



  • SuspiciousCarrot78@aussie.zonetoSelfhostedRevisiting Rule #3
    link
    fedilink
    English
    arrow-up
    4
    ·
    edit-2
    2 days ago

    Disagree with the phrasing myself.

    Yes “please ensure …” is nominally more proactive but “be prepared to explain” should force a technical framing in the first place, because the poster in on notice they may be challenged. Don’t write slop, read the room etc. It also explains why the rule exists / what the fail state is.

    The first one just sounds like so much corpo HR double speak.

    Be bold. Say what you mean - “We talk shop here. Be prepared to put up or shut up, else bonk”.

    If the user can justify their question (if/when queried), it doesn’t matter whether the question is hardware, software or other - it will have a grain.

    Additionally it might give people an foothold into OPs thinking process / whether they selected the right room in the first place.

    My $0.02 and YMMV








  • For $10, that’s quite impressive. I’m familiar with several of those games, including Just Cause 2, which I have run on i7-4785t (iGPU only) at 630p, hitting 69fps (albeit it AA off etc, AF 4x, textures high, shadows low, medium water and object details etc).

    OTOH, the i7-4785T doesn’t hit 90+ degrees to do it :) (sits around 70) and sips ~40W.

    EDIT: from the video, the test rig was -

    Test Bench Specs:

    • CPU: Intel Core 2 Duo E7500 @ 3.8GHz OC
    • RAM: 8GB DDR2
    • GPU: ATI Radeon HD 3850 512MB
    • Storage: 1TB SATA SSD
    • PSU: 450W 80+ Bronze
    • OS: Windows 7 x64

  • SuspiciousCarrot78@aussie.zonetoSelfhostedContinuwuity
    link
    fedilink
    English
    arrow-up
    12
    arrow-down
    3
    ·
    edit-2
    4 days ago

    Funny, I’ve had the exact opposite experience reddit vs lemmy. People are much more supportive here.

    Reddit is especially a shithole for that exact use case (sharing tech projects), because (I suspect) pecking order games and performing for imaginary audiences. Between that and the slop…yeah, nah. Fuck Reddit.

    YMMV