Live data from Hacker News

Claude Code does our releases now

aluxian.com

51–60 of 81 posts

Re: Claude Code does our releases now

#51
post #37
post #22

Earlier quoted context omitted.

We are talking about a Gen AI startup that has a handful of employees here. They have little excuse not to implement CI/CD, unless they lack confidence in their product's quality.

Ignoring the Claude AI release, why does releasing 1 - 3 per week implies low confidence and quality issues?

Because they should be spending one day writing scripts and github actions for their CI/CD system before pushing out new code by hand or AI assistance several times a week.

Releasing 1 - 3 times a week means it's 1 - 3 times more important to have a deterministic release process than if you release 1 time a week.

Re: Claude Code does our releases now

#52
post #42
post #16

Earlier quoted context omitted.

Why would a startup CTO choose to deploy a couple of times a week with a Claude script instead of implementing basic CI/CD? He makes it sound like he's running a clusterfuck and he isn't qualified for the job.

Aren't you assuming a bit much here? They have implemented CI/CD; their Claude script triggers the git branch part of the release flow, which is what in turn triggers the CI/CD pipeline—a completely normal thing to do, even in established teams (save for the Claude part). The guy automated the toil away using AI. Not that I would feel confident automating away that part in particular, but it does neither speak badly…

Because there is more toil in writing a blog post about how you use AI to do a release, than there is in writing a script that makes the AI and the blog post unnecessary.

But then you wouldn't have something about using AI to blog about.

Re: Claude Code does our releases now

#53
post #16

Earlier quoted context omitted.

> but your release cadence implies low confidence and quality issues. curious why would it imply that.

Why would a startup CTO choose to deploy a couple of times a week with a Claude script instead of implementing basic CI/CD? He makes it sound like he's running a clusterfuck and he isn't qualified for the job.

It really depends on the product. Especially if 0-downtime deployment is not possible.

Re: Claude Code does our releases now

#54
post #50

Is this not a bash script, but run through a "maybe it won't work this time" randomizer? Sometimes I feel like I live on another planet. Wouldn't you at least get Claude to write the bash script, confirm it works how you like, then run that? Why get an LLM to guess how to do it each and every time? At least they are still manually approving, which the title made sound like something they'd move on from.

We're moving fast from "we do this with AI because it's useful" to "we do this with AI because it's cool despite the fact that it's slower, more expensive, and non-deterministic"

It's like the "we made over our product in order to use technology X because it's cool and modern" that we've seen multiple times with node, go, rust, k8s, blockchain, .

Re: Claude Code does our releases now

#55
post #11

> So please No one has ever said ”please” to me in release instructions. Not that I miss it, I would find it weird.

It's a well known prompt trick to get better results. Scott Hanselman explained it like "You get to the nice corner of the internet."

But what if the LLM senses you're using the word "please" in the passive aggressive, patronizing, ironic sense? ;)

Re: Claude Code does our releases now

#57
post #24

This is cool and all, but I don't see why this can't be accomplished (and more deterministically) with a bunch of bash/python scripts. I've seen that done and work well in several firms for decades.

Agreed. This prompt is very easy to satisfy with 1970s technology.

JCL! My mom could do that. Fortunately she's safely and blissfully retired, so AI can't take her jerbs.

https://en.wikipedia.org/wiki/Job_Control_Language

>"the worst computer programming language ever devised by anybody, anywhere" -Fred Brooks

(Obviously when he said that, Fred hadn't been exposed to bash yet!)

"Please write a JCL script to orchestrate the release of my AI startup company's AI powered shop assistant from github. Did I mention that we use AI?"

    //*********************************************************************
    //*  M O L I N A I   R E L E A S E   P I P E L I N E                  *
    //*                                                                   *
    //*  Product  : AVATAR-SHOP-ASSISTANT                                 *
    //*  Company  : Aluxian / Molin AI                                    *
    //*  Purpose  : Tag, publish and attach artefacts to GitHub           *
    //*********************************************************************
    //AVTREL  JOB (AC),'MOLIN AI',CLASS=A,MSGCLASS=X,JOBPARM=TIME=1440,TYPRUN=HOLD
    //         REGION=0M
    //SET PRJ=MOLINAI,APP=AVATAR,TAG=v3.1.0,TOK=ghp_FAKE,REPO=aluxian/avatar
    //SET WS=/opt/molin/ws
    //MUSE PROC GIT=/usr/lpp/git4z/bin/git4z,CURL=/usr/lpp/bear/bin/curlbear,SHA=
    //ENV  SET PATH=/usr/lpp/git4z/bin:/bin HOME=/home/molin
    //SHA  EXEC PGM=BPXBATCH,PARM='PGM &GIT rev-parse HEAD'
    //STDOUT DD DSN=&&S,UNIT=SYSDA,SPACE=(CYL,1),DISP=(,PASS)
    //IFTAG IF (&SHA ¬= ' ') THEN
    //TAG  EXEC PGM=BPXBATCH,PARM='PGM &GIT tag -a &TAG &SHA -m &TAG'
    // ELSE
    //TAG  EXEC PGM=BPXBATCH,PARM='PGM &GIT tag -a &TAG -m &TAG'
    // ENDIF
    //PUSH EXEC PGM=BPXBATCH,PARM='PGM &GIT push origin &TAG',COND=(4,LT)
    //JSON EXEC PGM=IEBGENER
    //SYSUT1 DD *
    {"tag_name":"&TAG","name":"Avatar &TAG","generate_release_notes":true}
    /*
    //SYSUT2 DD DISP=(NEW,CATLG),DSN=&&PAY(0),UNIT=SYSDA,SPACE=(TRK,1)
    //SYSIN  DD DUMMY
    //ESDS EXEC PGM=IDCAMS
    //SYSPRINT DD SYSOUT=*
    //SYSIN DD *
    DEFINE CLUSTER(NAME(&&E) RECORDSIZE(255 255) KEYS(1 0) REUSE SPEED) -
      DATA(NAME(&&E.DATA))
    REPRO INFILE(P) OUTDATASET(&&E.DATA)
    IF MAXCC>4 THEN SET MAXCC=0
    /*
    //P DD DISP=SHR,DSN=&&PAY(0)
    //POST EXEC PGM=BPXBATCH,
    // PARM='PGM &CURL -sS -X POST https://api.github.com/repos/&REPO./releases
    //      -H Authorization:\ Bearer\ &TOK -H Accept:\ application/vnd.github+json
    //      --data-binary\ @DD:EJSON'
    //EJSON DD DISP=SHR,DSN=&&E.DATA
    //STDOUT DD DSN=&&R,UNIT=SYSDA,SPACE=(TRK,1),DISP=(,PASS)
    //RID EXEC PGM=IKJEFT01
    //SYSTSPRT DD DSN=&&ID,UNIT=SYSDA,SPACE=(TRK,1),DISP=(,PASS)
    //SYSTSIN DD *
      alloc fi(r) da('&&R') shr
      execio * diskr r (stem x. finis
      do i=1 to x.0
        if pos('"id"',x.i)>0 then parse var x.i '"id":' id ',' .
        if id¬='' then say strip(id); leave
      end
    /*
    //* BLDL fix — after the ‘Q3-FY24 LLA purge’ when an ill-advised UPDATE to
    //*   SYS1.LINKLIB nuked the BLDLRES chain, froze LLA, and threw half the
    //*   plex into S047s until Ops spent nine hours baby-sitting a cold-start IPL.
    //COPY EXEC PGM=IEBCOPY
    //SYSUT1 DD DISP=SHR,DSN=&&R
    //SYSUT2 DD DSN=&PRJ..G.&APP..R(&TAG),DISP=(NEW,CATLG),
    //             SPACE=(CYL,(1,1,1)),DSORG=PO
    //UPL EXEC PGM=BPXBATCH,
    // PARM='PGM &CURL -sS -X POST
    //      https://uploads.github.com/repos/&REPO./releases/$(cat &&ID)/assets?name=shop.jar
    //      -H Authorization:\ Bearer\ &TOK -H Content-Type:\ application/java-archive
    //      --data-binary\ @&WS/&APP/dist/shop.jar'
    //STDENV DD DUMMY
    //PEND
    //GO EXEC MUSE

Re: Claude Code does our releases now

#58
post #50

Is this not a bash script, but run through a "maybe it won't work this time" randomizer? Sometimes I feel like I live on another planet. Wouldn't you at least get Claude to write the bash script, confirm it works how you like, then run that? Why get an LLM to guess how to do it each and every time? At least they are still manually approving, which the title made sound like something they'd move on from.

And the bash script would be shorter and easier to write than the prompt.

Re: Claude Code does our releases now

#59
post #18

Earlier quoted context omitted.

>can we pls ban these coders from the industry This is the main problem with AI, we can no longer distinguish bad coders/managers because they use AI heavily and which can say knowledgeable things at times while having no idea on what they are doing. I now say things that will not make sense to a good coders (like saying the steps out of order), to find out such people.

> ...and which can say knowledgeable things at times while having no idea on what they are doing This has been a thing forever, before AI. Lots of "programmers" at BigCorp have hacked the system and they know how to say the right words relevant to BigCorp's tech stack. Knowing how to actually program is entirely optional. (You can just copy-paste boilerplate from BigCorp's monorepo, writing boilerplate will be your d…

yea for sure they hacked the system for promotions but they weren't checking in AI slop for rest of us to deal with.
Post reply on HN