656 karma · joined November 30, 2011
Now all you really have to do is chat with claude about what you're thinking about building.
In plan mode, claude can't edit anything and has some extra "you're an expert at planning!" prompts prepended to your initial message in plan mode.
And then either when you're ready or Claude thinks the "plan" is gelling, it'll suggest stopping and let it write up a detailed plan. CC will dispatch some "planning agents" with prompts that your 'main' CC has crafted for that agent to plan for within the context of your conversation and what parts of the codebase it should look to integrate/explore.
Once all that is done, it will display it to you and then offer to "clear context and implement" - where it will just get to work. Or it will offer to go back to chatting and resolve whatever was misunderstood, or if you had a new idea you wanted to mix in.
These plans are saved as markdown in your .claude/plans directory.
Plan mode is handy on the on-off. But if you enter another plan mode, thinking claude would learn from, or build off a previous plan spec, it won't unless you explicitly say something like "read previous plan <path to plan file> and re-use the scaffolding directives for this new project"
Like most spec driven development tools, GSD works well for greenfield or first few rounds of “compound engineering.” However, like all others, the project gets too big and GSD can’t manage to deliver working code reliably.
Agents working GSD plans will start leaving orphans all over, it won’t wire them up properly because verification stages use simple lexical tools to search code for implementation facts. I tried giving GSD some ast aware tools but good luck getting Claude to reliably use them.
Ultimately I put GSD back on the shelf and developed my own “property graph” based planner that is closer to Claude “plan mode” but the design SOT is structured properties and not markdown. My system will generate docs from the graph as user docs. Agents only get tasked as my “graph” closes nodes and re-sorts around invariants, then agents are tasked directly.
I've been doing this for decades, and I've never seen that attitude work with any 'leader.' I'd hate to work for you. Ever hear of servant leadership? Or hear the line "My job is to clear the runway for you"
Managers are cost centers, 'your' employees are what keep you employed, give them the respect they deserve.
Tell me more about how humans are criminals for existing? You hate the idea of someone standing on your piece of dirt without your express say-so that your preference is to live under an umbrella of fear and suspicion of your neighbors?
Forgive me, this is not how I want to live. I wonder how many Norwegians who settled in Minnesota were illegal immigrants?
I guess I should thank you for saving my time? Plenty of others in this space.
I have since solved this problem: 1) Tool-Check PLUS from Wera: fantastic and highly condensed set of sockets, bits, ratchet and driver. all high quality. 2) Leatherman ARC: replaced a more basic multi-tool. I like the high quality blade material and the whole thing is very durable, i can squeeze and crank the pliers without worry.
The leatherman is a bit of a splurge but I highly recommend the tool-check for tool-using EDC minded folks. They make good gifts too.
sidebar: I've had a lot of problems with the AppleTV plex app, it does not play some HDR content well, or at all. Plex on my nVidia Shield works well.
I really wish I still had that thing.
If I wanted a dual motor config I'd just have ordered a Ford, which would probably deliver sooner, for less money, and probably better supported in the long run.
I am willing to wait but Rivian and the others have lost so much credibility to deliver that I do not even trust a deposit with them.
So I will continue to wait until manufacturers can ship these things as ordered.
If I am correct, the hiring manager knows you have 4 years experience and isn't necessarily focused on hiring "Senior" but more interested in people who have a proven track record, which 4 years in is.
The way to really distinguish my guess from reality is to ask a lot of questions about the expected duties. If they fall within or near your current experience then you're all set.
If they are asking you to make big decisions about technical direction or influencing development teams then the hiring manager is really looking for a more experience person. Still, in my opinion, if they're giving you an interview and if you think you can meet the expectations then no harm, otherwise you should pass.
I have a M1 Max and have mostly avoided running SD on my machine since many SD implementations are built for windows or linux, supporting apple silicon had been less of a priority. Many of these frameworks will work on an M1 but i find using colab is simpler with similar performance.
However, since apple has released[1] Core-ML support SD models I have been trying some native apps like Mochi-diffusion[2] they seem to work well but are not yet as feature rich as the older webuis
1: https://github.com/apple/ml-stable-diffusion 2: https://github.com/godly-devotion/mochi-diffusion
I have always thought that our digital world should service humanity, meaning I want software that augments my human abilities, I extract value from it vs how things are now where services only exist to monetize human activities.
For example: I want an AI mind that helps me discover and use information while factoring in my personal biases, morals and needs.
I know that reeks of idealism but it is what it is.
Meanwhile, we have stable diffusion which hits close to the mark. One of my hobbies is drawing and painting. I don't have a lot of time to commit to it so I try to be sure of the concept before I work on it. I find this to limit my creativity a lot, I get stymied on an idea "is this worth 40 hours of precious free time?"
With SD I am able to quickly iterate on concepts. Of course, its not giving me exactly what I have in mind but it gets close enough that I either renew inspiration or I see my concept in a new light and move on.
The fact that I can self host, tune and even retrain SD gives is super attractive. I sincerely hope more open source models like SD get released so I can build an personal AI ecosystem that meets my needs.
I plan to install a huge solar array with a battery house. I'd like to run everything off electric, including the heat.
I am in early days of thinking about this and I have time to plan. Anyone have insights on electric heat in ultra cold environments? I assume I can simply scale up a solar array and battery capacity to meet needs (dead of winter, with spans of cloudy days). The only fossil fuels I want on property are for equipment and if I must have it, a backup NG generator.
I don't know anyone who runs heat pumps in MN, I'm sure there are some but most folks are burning NG or wood pellets. Electric heat seems relegated to secondary needs, like base boards or heated floors.
*edit cloudy
I don't know when this was created but I wonder if a new StableDiffusion (w/dreambooth) powered version would be more interesting since the observer/editor gives their interpretation of "female gaze" for the subject VS what the author tells what agency is. It would be illustrative to see where the public matches or deviates from the author's perspective.
Two major pain points for me:
1) I can't play my whole library. During the beta I was able to play pretty much my whole steam library but that changed after GFN went live. I accept that some titles may have technical limitations but thats far from universal. Like, why can't I play Red Dead Redemption II on GFN but I can on Stadia? Pure publisher bullshit, that's why
2) I wish Nvidia made it possible to apply mods to games. Steam workshop mods work if supported but the majority of games do not support that. I don't often mod but when I do its usually to correct some 'quality of life' issue.
I ended up buying a Steamdeck to game on instead of GFN.. I hope the arbitrary restrictions holding it back disappear someday
It is ridiculous to question if there is any effort expended.