Sabitlenmiş Tweet
👋 I'm Jonathan – PM at @MOIAmobility, indie dev at night. AI tools got me back into building apps alongside my PM work. Here I'm going to share what I build and learn.
English
Jonathan Glasmeyer
515 posts
Jonathan Glasmeyer
@jonathan_builds
Senior PM @MOIAmobility by day, building & shipping solo at night
Hamburg Katılım Şubat 2022
445 Takip Edilen147 Takipçiler
But it's not included in plan? Prohibitively expensive, no?
Quote from docs:
"Fast mode usage draws directly from usage credits, even if you have remaining usage on your plan. This means fast mode tokens do not count against your plan’s included usage and are charged at the fast mode rate from the first token."
English
puuuumped. though as far as i can tell, there's no way to leave it on. it basically turns off after every turn. at least in claude code desktop.
ClaudeDevs@ClaudeDevs
Fast mode now defaults to Opus 4.7 in Claude Code. Try it out today with /fast
English
Have you run into CCD spawning new sessions by itself? Shows as a popover suggesting prompt for new session. Starts in a new work tree. Freaking love it. "Spawn new CCD sessions for these 5 bug reports, each".
It's superior UX-wise to the terminal sub-agent spawn / teams of agents stuff, I think.
English
been giving the claude code desktop app a proper run the past 48 hours and i'm LIVID.
it's just so much better than i thought it'd be. 😭
i also don't know what they're doing with tokens in it but i've been running at full throttle and never drop below 80% session limit.
i can't handle the gpt 5.5 vibes for coding. as much as i don't trust anthropic as a company, i still trust opus far more than gpt as my primary driver.
(i know i know, that's not everyone's experience. at the end of the day every single thing in this space is subjective.)
English
@Shpigford Claude Desktop Code with worktrees & setup hooks work well for me, switched 3-4 weeks ago from terminal and haven't looked back
English
bummer of the year (which is just a day in the AI world).
this is the nail in the coffin for me with anthropic.
conductor is arguably as big of a productivity boost as anthropic is (was).
time to port all my skills/workflows over to gpt 5.5. (still in conductor!)
Charlie Holtz@charlieholtz
Here's what Anthropic pricing updates mean for Conductor users: - You can officially use your Claude sub with Conductor - If you're on a max subscription you get $200 in credits and then can pay at API costs - If you use Big Terminal Mode you won't be affected We're going to keep building the best interface for the best coding agents! Excited to show you what we've been cooking🫡
English
@Shpigford Frankly don't understand the fuss/hate on Anthropic. They've now encoded their position on 3rd party & programmatic use now very clear, imo. Plus, 50% more weekly use until mid July?! come on now :)
English
i know there's going to be an adjustment period but man do i not enjoy gpt as my primary model.
100% subjective in every way.
just been a real bummer of a morning using gpt for everything. i'd become so unbelievably productive and so i acutely feel when something changes.
English
@amorriscode @raroque @bcherny That's probably the feature I looked forward the most since the CCD relaunch 🫶
English
I asked @bcherny what his favorite claude code feature was
he said it’s remote control
BUT not using “/remote-control” 👀
Learned something new!
English
@Shpigford i'm just using gdrive, mirrored to my mac hd, so claude can do all the organization/folder taxonomy etc
English
does no one deal with decades of PDF files? why is everyone recommending obsidian?
i don't need text editing. i need file storage with an substantial OCR + indexing + search layer.
there's basically *zero* native raw text here.
Josh Pigford@Shpigford
evernote is doubling the price of my subscription to an arguably insane price, even for a PE-acquired business. we've got 15+ years of docs stored there (huge % are scanned PDFs). trying to figure out my move here. i don't want another note taking app. i kind of just want a vault for dumping all our random document files (again, mostly PDFs) that's thoroughly indexed/searchable. big caveat: needs to be also be easily shared with my wife. wonder if there's something there from a biz perspective? 🤔
English
Opus 4.7 is wonderful for me in Claude Code. It seems to have taken some diligence & seriousness from GPT5.4, very thorough, sharp. Hasn't lost the Claude SOUL though I feel. Wonderful coding partner.
English
@noahzweben we're using github.com/anthropics/cla… quite a bit for similar use-cases. would you recommend migrating over?
English
Claude Code Routines are here! In addition to a schedule, you can now trigger templated agents via GitHub event or API – with our infra & your MCP+repos
They've changed how we do docs, backlog maintenance and more internally at Anthropic
Get started at claude.ai/code/routines
English
@Shpigford Same. Reliable code tasks spawned from Dispatch; inspectable in desktop app. It just works. Feels like Mac vs Linux ecosystems..
English
i turned off my 🦞
the past week of trying to force it to work with other models was a massive drain and i was able to replicate the vast majority of impactful functionality with a mix of claude cowork + code.
English
@michael_chomsky This is already nicely solved via Obsidian. Mine e.g. lives on Macmini, is accessible from both Openclaw & Claude Cowork (Dispatch)
English
Here’s an idea.md for anyone who isn’t scared to build an ambitious product:
Someone’s going to make 100-1000M dollars building a self-updating personal knowledge base that syncs with imessage, twitter, email, chatgpt/claude/claude code/codex messages.
This knowledge base will have an MCP to be accessible from anywhere. You’ll be able to edit it like notion, style it to your liking, and set rules about how data is organized.
Once it gets mature enough, it can even proactively suggest things that will improve your life, as it knows everything about you.
Unlike some memory systems, this will just be files so completely observable.
The closest thing to this that exists is Mintlify’s KB and Notion, but both are more for enterprises than individuals.
Just paste this into Claude Code, ask it to perform a socratic interview about ambiguities (or just use gstack), clank for 2 days, get Chamath and Karpathy as your first two customers, and do YC.
You’ll have competition at some point but it won’t matter because you’ll be better at UGC and paid ads.
Chamath Palihapitiya@chamath
This may be a dumb question but I’ll ask it here anyways: I can’t find a good way for my various AI chats to automatically sync its conversation history into a structured knowledge base. So that as I update various chats from time to time and refine context, my knowledge base automatically grows with this new info.
English
@ryancarson @DevinAI More expensive though huh? Running these openclaw codex chores with a pro plan, easily
English
I use @DevinAI to do a daily e2e automation that signs up my app, goes through onboarding, tests the chat functionality and all of its tools, and then sends me a report every morning. I can even watch the video.
I get this report every morning in Slack when I wake up. It's reassuring to know that core functionality in the app is working and if any bugs are found they get auto-fixed
English
@Shpigford Hate these empty phrases, and they an't be SOUL.md-ed out of the model..
English
GPT 5.4 CONSTANTLY says it's giving me an "honest take" and that it's going to be "blunt". almost every reply. it's absurd.
Barron Roth@iamBarronRoth
@Shpigford how can i get JUST the “honest take”
English
@kimmonismus Didn't the leak only hint at Mythos, the one being way to expensive to serve to regular plans? 🤔
English
@subham_agrawal_ @Shpigford i just have at least 3-4 distinct work & personal contexts. combined w/ not being able to `/new`/`/clear` on the dispatch session, dumping all of these in a long-running context is pure context poison.
English
@Shpigford @jonathan_builds Hot take: Dispatch's single thread FORCES focus. I avoid context switching & get WAY more done. Am I the only weirdo who loves it? 😅
English
what i want out of a personal ai system...
* dynamic memory
* plugins (for interacting w/ APIs)
* skills (for running common workflows)
* scheduling
* remote access
* chat threading/topics
and i want it all powered by opus. and i don't want to pay $50-100/day for it. 🫣
English
@Shpigford i guess it's the mvp to copying the essence of openclaw, which is this one agent to dispatch jobs..
hoping for a more elaborate multiple agents setup that integrates better w/ cowork&code projects at some point.
English
@jonathan_builds I really do like Cowork a lot. I find Dispatch to be...kind of a nothing burger. The single-thread for everything kinda kills it for me.
English
phew.. spent some hours porting openclaw crons/heartbeat actions & skills to cowork/dispatch. i must say i _do_ like the first-class-citizen ui for inspecting the setup.
English