@jonochocki Two months of burning the week in 3 days is brutal. I stopped guessing and started wanting a live spend meter + pause mid-response when it's looping
@DannielDevOps Quiet 7% from tighter prompts + selective tools is the kind of win I actually care about. Session spend you can see beats opaque "it felt cheaper"
@jimmy_longbow_ 200:1 in/out and the bill is prompt processing — yeah. I want that meter on screen during the run, not a surprise later. Desktop agent + session spend reporting so I can steer
@hars_7086 Score and token burn in the same screenshot is exactly right. Better agent number + 3x the output tokens is a worse bill if you leave it unsupervised. I watch session spend and pause when the burn looks wrong
@vernons Yeah — plan cancelled because the tokens ran out before the work did. I want session spend visible while the agent is still running so I can pause mid-response instead of finding out on the invoice
Tokens burned out before the work was done.
I'd rather see the spend meter while the agent is still typing — so I can pause mid-response when the burn looks wrong — than cancel a plan after the invoice.
Desktop agent. BYOK or bring my OpenAI/Grok sub. Session spend on screen. Keys in the OS vault.
@charleyecom Context so small it forgets everything mid-thought is exactly why I want threads that never compact. Small window + aggressive summary = stupid agent. Keep the history, browse it
@skyw4Ik3r@AnthropicAI Free reset → gone in 2 hours because it spawned multi-agent reviews you didn't ask for. I don't fire-and-forget overnight runs. Steer, monitor, pause — or watch the budget vanish
@AhmedReza 15M tokens on a straight design audit and it reruns the same workflow?? That's the mid-session tax. I'd rather pause mid-response and steer than let the harness burn itself in a loop
@roeishous@RLanceMartin 154 of 219 as a bare name is brutal. That's not "skills," that's a name tag. I want the full context still there — inspect it, don't summarize it away past 1%
@aithren_aj Yeah — polite bullet summaries that quietly delete the one detail you needed. I've stopped trusting compaction mid-session. Prefer a local agent that never compact + a context browser so I can pin/trim myself
Compaction just dropped the one detail that mattered.
I’d rather keep the full thread on a local desktop agent — never compact — and browse/pin/trim context myself.
Bring my OpenAI/Grok sub or BYOK. Session spend visible. Pause mid-response when it loops.
@Riks_6@ClaudeDevs 22 hours after weekly reset and x20 is already gone. That's not a plan, that's a timer. Desktop agent + BYOK/bring-sub so I'm not racing the reset clock first thing.
@Kirjakulov@AnthropicAI Free reset is nice. Cutting the flow mid-thought still sucks. That's the morning wall — I'd rather steer a local agent on my own keys/sub than wait on a provider reset.
@cheeoonai $1,227 and you're only at 38%. Concrete math. I want that visibility on the box while I work — session spend + bring-your-sub, not a post-mortem on a weekly bar.
@DonaldChitana Forgotten how fast the $20 bucket dies mid-coding — then Max hits too. Morning quota babysitting. We run on your OpenAI/Grok sub or BYOK so the wall isn't a surprise plan reset.
@AnthropicWay That morning reset → half-week-gone hit is exactly the usage wall I hate waking up to. I'd rather bring my own sub on a desktop agent and watch the burn, not babysit opaque weekly bars.
Woke up to another usage wall.
I'd rather bring my own OpenAI / Grok sub (or BYOK) on a desktop agent than babysit opaque five-hour and weekly bars first thing.
@CristianXIV One Astra prompt for a whole weekly bucket is exactly why I want the meter before prompt #2. If the burn rate is visible while I steer, I can make the tradeoff before the quota is gone.
@SilasHelyfts@shownotover 36% from one Chat is the invisible-meter problem in one number. I want to see session spend while I'm still steering, not discover the limit after a single conversation.
2 Followers 23 FollowingFounder. Giving AI coding agents the context that lives in your team's heads. I measure what Claude Code is actually told. Tel Aviv.
94 Followers 732 FollowingI design brands and motion, and build practical AI workflows. Solopreneur since 2022. Gondo Creative Studios, Zimbabwe. Creative work and systems consulting.
827 Followers 2K FollowingAI Research Engineer • Cybernetic Systems • Creator of Cortex, a constitutional metacognitive runtime for persistent coding agents. https://t.co/fkMwd62Z7z
194 Followers 338 FollowingLicensed portfolio manager, developer of trading systems and now building AI infrastructures for financial institutions.https://t.co/vgb0Lkh8Vs
811 Followers 1K FollowingLazy investor who is building everyday tools (https://t.co/BTfJ32tmn5) and a SaaS for SG real estate (https://t.co/HHDRIyBkgW).
8K Followers 7K FollowingDark trap/experimental producer. Custom beats, sample packs, and sovereign workflow for artists who don't sound like everyone else.
199 Followers 153 FollowingBuilding @SpliceRun. Self-host n8n, Hermes Agent and more on a server you own. We keep it updated, backed up and alive. First 100 get 6 months free, no card.
100 Followers 259 FollowingYour coding agent knows the code. Give it the robot run. Investigate failures and check changes with Cerulion.
Backed by @ycombinator
440 Followers 952 FollowingAI employees for small businesses — receptionists answering every visitor 24/7, in any language. Local SEO to get you found on Google Maps. https://t.co/zuE0qWcUIi
173 Followers 255 FollowingI quit my job to build apps alone.
Turns out shipping was the easy part.
📱 Split Today · live on Google Play
📊 Mondays: real numbers, what I got wrong
90 Followers 427 FollowingI spend a lot of time on AI, agents, developer tools and how software is changing. I share what I’m building, learning and finding useful along the way.
14 Followers 28 FollowingiOS developer.
Promoted into meetings, still living in the terminal.
Model evals, GitHub picks, and notes from actually shipping it.
51 Followers 287 FollowingFull-stack developer building with AI.
I write about AI models, agents, and tools. What's new, why it matters, and the details behind.
I want 100 followers.
2K Followers 701 FollowingFounder & CEO @Yobi_App — building AI employees for small businesses | https://t.co/oPkS9l3pvl | Ex-NASA, Milken Scholar, Cornell CS | Also built @TrepHub, Call Sumo
2 Followers 23 FollowingFounder. Giving AI coding agents the context that lives in your team's heads. I measure what Claude Code is actually told. Tel Aviv.
94 Followers 732 FollowingI design brands and motion, and build practical AI workflows. Solopreneur since 2022. Gondo Creative Studios, Zimbabwe. Creative work and systems consulting.
827 Followers 2K FollowingAI Research Engineer • Cybernetic Systems • Creator of Cortex, a constitutional metacognitive runtime for persistent coding agents. https://t.co/fkMwd62Z7z
1K Followers 382 FollowingShorted frauds until Bill Hwang stole all my money. Doing my best to embrace facts in a world that's become all about perception.
102 Followers 210 Following15+ yrs shipping software at scale — Apple → Salesforce → Drones | Head of Eng | LinkedIn Learning Instructor | Builder of open-source dev tools
12K Followers 4K FollowingBuilding auditable coordination systems for AI agents: provenance, abstention, retraction and resilient handoffs. Creator of CLE. Evidence over hype.
512 Followers 778 FollowingWordPress 6.0+ UX Developer | Gutenberg | Block Building | Dev Tooling | React | Building websites and PWAs
Available for consulting and development.
517 Followers 682 FollowingYou can just pretend to be successful (unemployed, .125x dev, 6’1” btw.). Extract Markdown & LaTeX using OCR: https://t.co/GgavTVMVFH (€6 quadrillion MRR)
670 Followers 822 FollowingI'm a computer engineer specializing in custom software and AI, management systems, CRM, ERP, mobile apps, digital newspapers, and e-commerce.
815 Followers 337 FollowingAI Engineer building agentic harnesses & inference optimizations. I write about interesting things in AI. ex-DoD Intelligence Operator; ex-DraftKings;