Pocster

You know what. If someone said that title too me I'd be well happy.

“Programming “ 2026 style

Currently ha is just doing time of use as the cloud rubbish was attempting to do before . Later I’ll make it based on use not time .

I was (expletive deleted)ed off again ! SE cloud constantly failing to execute battery schedule. So bit the bullet and with my best mate ( chat ) got it working in home assistant. At the moment still time based - but working !!! Then we can get far more advanced as to when it discharges specifically to Tesla ! Awesome stuff ! I owe chat a good few drinks !!

Lol, just get to scout and suggest no code changes.

Lol. Mines a hobby not "real" like yours. But my issue even with previous projects and chat was lack of repo access. Go on give claude repo access just to look/report. BE BRAVE!

Chat still ignores things which it gets in it's prompt still! The 'loop' is everything goes back to chat. Ledger, patch outcome, scout etc etc etc. I'm just human doing boring stuff in the middle that creates project but writes zero code. Because my ambition was i do ZERO code we need a tight and rather repetitive loop. My assumption is that Claude nails this in a far better way. But chat with pi as repo scout just requires drag n drop of output files into chatgpt window. I was thrown by codex thinking it was chat but with repo access. But it isn't. It can patch etc but it doesnt understand the project and its structure. ChatGPT does - proves it constantly even if it has the odd wobble now and again. @SimonD I'm confused what the issue is TBH. Claude see's the repo thats the best bit. I have to use local llm to scout repo to uload to chat. Also I naturally assume Claude better reasoner and coder than chat. So I'd assume this arrangement would work better tbh. After all I'm doing a 'poor mans' claude code

LOL! I understand exactly what you mean and exactly that bug. But as I've said to date not had to look at 1 line of code. Chat provides the patch but we have a array of multiple tests it has to pass. If it fails the tests etc I get an output file of why. This is then chucked back at chat where it either does another patch or a better scout. I've added stuff to the pi harness to reduce failing patches quite substantially e.g. pi can now grep/sed/tail files into its prompt for local 120b scout. Commits only happen when I say so. Ledger updates must match i.e. what we did, what was tested, what was commited. Basically so many guardrails that any patching applied has to be rock solid. I spend 90% of my time after supplying the spec/task pressing '7' to scout or '2' to apply patch. A few copy n pastes when chat gets pissy and stuck i.e. it tells me what it needs and I do it manually. TBH it's working really well. Soon I'll tie this into the 1st version of voice transcript and upload a demo. Here's where I spend most my time!

Awesomeness! Refined the harness code ( I.e told chat it was shit ) - patches approval really high now . Implemented but not ‘ voice ‘ tested … ” play Coldplay “ ” play head full of dreams “ ” play that Coldplay album I played last Tuesday “ ” what Coldplay albums do I have? “ ” play something by Coldplay “ etc etc nice and broad . Also if you mis say or more likely it mis heard “ foldplay” will ask if you meant “ Coldplay “ Again not linked into the whisper transcript yet - just fuzzy qwen:3b ‘guessing’ based on intent and what’s in my LMS database . Currently adding Postgres database functionality

Buy it ! Prove me wrong ! Said that to a girl a few months ago … didn’t end well !

My link is!. water in, electric in, drain out. Cost is upsetting though.... Looking back at cheap mobile units where you jsut sticka vent out the door and seal that gap up....

I can only think of 1 reason someone would gag............

But now I'm thinking... https://www.argos.co.uk/product/7623899?deeplink=true&gclsrc=aw.ds&&_$ja=tsid:11675|acid:542-316-8335|cid:6514064352|agid:76370554737|tid:kwl-3500001|crid:610242503975|nw:g|rnd:4608841013949138266|dvc:c|adp:|mt:a|loc:9045627&utm_source=Google&utm_medium=cpc&utm_campaign=6514064352&utm_term=&utm_content=text&utm_custom1=76370554737&utm_custom2=542-316-8335&cmpid=GG05X&GPDP=true&gad_source=1&gad_campaignid=6514064352&gbraid=0AAAAAD9II9kB4EdwoWJiJtjRBVlI1afcv&gclid=EAIaIQobChMI4sjlieLZlAMVhotQBh3GDSczEAAYASAAEgKcafD_BwE £400 and make a 'panel' so the vent isnt just hanging out a window/sliding door.

https://cooleasy.co.uk/products/powrmatic-vision-h2o???? 3k! presumably is the condenser aswell???

Don't want someone looking! DIY please

Water cooled air con? As far as I understand no vent through wall. Air con unit just needs pipe work to condensor unit. I could mount that under MVHR unit and share its drainage channel to outdoors???

OK! Back here again! So need to cool 2 bedrooms with AC. No floor standing shite. So I guess single contained units; 1 in each room no split system. Going to be a BASTARD to install i.e. to get the vent pipe through. (TWICE!) Am I allowed to vent into the loft? 🙃 has it GOT to be outdoors? Electrics ok to wire off either a plug or off an existing socket ring?. No chance getting wiring from AC back to CU.... I'm under SWMBO orders to investigate!

Then you get sweet moments where 15 patches all go straight through 😎

Why not! Its hot! I'm in the garden fighting chat. I actually got it to say "I cant do it we need a coder with repo access". I beat the bitch into submission. We'd been trying and failing for 15 minutes. I asked it what it needed. Gave me a 'sed','tail','cat' answer. I did it manually in terminal uploaded it. 60 seconds later fixed. But yes "no coding skills" with an LLM is only going to get you toy stuff. Remember OpenAi said 70% of chat is written by chat and they expect that to be 100% soon. So the SE/coder role now changes to LLM manager/architect/controller. But yes ; I assume even Claude can get it's knickers in a twist! But from how I was using chat just 12 months ago (and its capability then ) to today is phenomenal ! We go do rabbit holes and poor design decisions. Sometimes it does things I didn't ask and when I find out later I tell it where to go! But! - what I am achieving in terms of work flow with zero human coding is pretty magical even if not completely local yet. Imagine where it will all be just 12 months from now! Oh! , I can hear the clink of ice in a G n T .......

Bit like giving an 8 yr old a chainsaw and then telling them to chop trees….

2 days of futility . yes Claude does it all Been fighting ! minstral in pi harness (expletive deleted) me slower than Daniel on a Friday behind bike shed . Swapped back to 120b and qwen next coder - another folly . Chat just picks apart its patch suggestions that no matter how we change the harness / structure prompts it ain’t happening . More chance of Stella saying yes . 3rd time she usually does but always wants a couple of quid extra . So ! I am back where I started ! Chat does prompt . 120b does scout / repo . Chat then does patch . Added a ledger ( think Claude does similar ) so what’s been done / what’s outstanding all recorded so I don’t get the dementia chat next day saying “ xxx needs doing “ when we’ve already done it . I’ve written more abuse at ChatGPT in the last 3 days than probably 3 yrs of posting on here . I *think* now we have a workable system . Local llm doing reasoning and or patch generation not good enough yet . So now finally rather than (expletive deleted)ing about on tools I’m back on the actual project . Sometimes you need a sore hand to learn not to touch it .

So slow ! Like a bitch with a chain around her ankle chasing you for payment

Minstral 3.5 !!! could be scout and patch constructor with gpt5.5 just confirming or doing new prompt !!! local llms move so fast !

Going amazingly well! Swapped to pi harness with custom code around it. Codex was dropping connection and being a dog - local YEAH! BS Pi is 100% local. So chat produces a scout prompt. I run it with 1 keypress from custom harness and pi. Upload back to chat if it passed/failed all the tests. If it's good chat does patch. Another single menu prompt and patch applied. Upload back to to chat if passed/failed. This loop is going REAL good. Took a few days to get the harness right with various models and bugs Treating chat as software engineer/architect is the way to go. Alexa's timer feature for example basically added in under 1 hour. - a good test!. Not touched or even look at 1 line of code.

Rabbit hole of rabbit holes! Tried most models ollama and mlx. Annoyingly things like DeepSeek cant do tools!! - just a friggin chatbot! codex,pi, continue cli, our own harness. Round and round we go! Getting largely no where! So! Conclusion! No local LLM is capable of producing the code and understanding my repo. Created a new harness using codex which produces an evidence pack that I uplaod to chat. Chat then does a patch. The new harness has a simple menu to save infinite copy n paste. Unsurprisingly codex works best with OPenAi models.... whom would of thought!