Declutter

819 posts

Declutter banner
Declutter

Declutter

@declutterKE

mostly AI, Gleam and Cloudflare. AI guy @moneyacademyKE

Katılım Şubat 2015
627 Takip Edilen118 Takipçiler
Sabitlenmiş Tweet
Declutter
Declutter@declutterKE·
Playbook.md Taste.md but for loops. Create a document that establishes a reusable design pattern and testing playbook for current and future projects to systematically prevent, diagnose and resolve common issues across various components. Use it to decide how loops are queued.
English
0
0
0
26
Declutter
Declutter@declutterKE·
The Unison programming language is a technological leap in terms of how we design software. Unfortunately it is bound to the Haskell ecosystem which is very annoying to most devs. Been thinking of rebuilding it in Gleam using Deepseek v4 pro max on CommandCode. Wish me luck
English
0
0
0
8
Declutter
Declutter@declutterKE·
@MrAhmadAwais defaulting to deepseek-v4-pro . its good enough for everything
English
0
0
0
7
Ahmad Awais
Ahmad Awais@MrAhmadAwais·
Command Code with open models and /design skill. 🔥
English
15
5
161
8.8K
Declutter
Declutter@declutterKE·
@samhogan whats the cost savings running glm5.2 on your stack compared to provider api
English
0
0
0
93
Sam Hogan 🇺🇸
Sam Hogan 🇺🇸@samhogan·
Looking like we'll 4x revenue in the next 90 days, thnx GLM 5.2
English
6
3
144
14.9K
Command Code
Command Code@CommandCodeAI·
We are doubling the usage limits to: 5-hour limit — 30% of your monthly credits Weekly limit — 60% of your monthly credits For all plans Go/Pro/Max. Provider and extra usage has no limit. We've also added major capacity, new providers, and a lot more GPUs to our self hosted infra. - Most models now run 5x to 8x faster - p50 TTFT: 5s → 965ms If Command Code felt slow the last few weeks, give it another go. It's fast now. Bigger plan dropping soon too. If you like the Go plan, this one's gonna hit home for y'all.
English
27
6
303
21.7K
Declutter
Declutter@declutterKE·
The entire online business model is dying. Internet marketing is dead. Bots don't click on ads.
English
0
0
0
10
Declutter retweetledi
WithoutHistory
WithoutHistory@WithoutHistory·
Mr Beast did poverty tourism while iShowSpeed showed the everyday Africa.
English
400
5.4K
43.7K
320.7K
Vipul Gupta
Vipul Gupta@vipulgupta2048·
shipping DX upgrades nonstop. like this one, review the full plan in the terminal before anything runs. todos on my list, 1. execute with permissions bypass 2. explain plan with diagrams/mockups 3. plan mode with your chosen model
Ahmad Awais@MrAhmadAwais

Command Code v0.34.0 is here! 3 new versions today, we're shipping fast!! lots of quality of live improvements. - feat: add Claude Fable 5 - fix: resolve [Text#N] paste placeholder into text for slash commands on submit - feat: major rehaul of lists and windowing /models /skills /mcp /agents - fix(tools): recover tool calls whose args are wrapped in a one-element array - feat: command aliases + consistent picker design across the TUI - fix: /context counts only enabled skills - fix: prevent flicker and OOM when scrolling long lists in /skills and /models - fix: a crashing MCP server no longer takes down the CLI - fix: stop running shell commands on Esc instead of waiting for them to finish (#428) - fix: /model and /configure-models pickers scroll within a viewport (selection bar, no OOM) - fix: /model selector scrolls within a viewport instead of jumping to the bottom - fix: /skills no longer flickers while scrolling or reprints a duplicate header on exit - fix: Per-task model overrides now fall back to ZDR-safe defaults on background tasks - fix: /context skill count and print mode now honor skills disabled via /skills - fix: write taste-relative paths so taste.md preferences actually save (Windows + macOS)

English
3
0
5
704
ℏεsam
ℏεsam@Hesamation·
POV: you realize top 1% Codex users run agents 71 hours/day in a 24-hour day. that is ON AVERAGE.
ℏεsam@Hesamation

OpenAI reports: Codex active users grew 5x in the first half of 2026. this is from OpenAI's paper released today, based on the data from Codex. main takeaways: > non-dev Codex usage is exploding. Since Aug 2025, it’s up ~189x in orgs and ~137x for individuals > inside OpenAI, Codex is now 99.8% of their output tokens and ChatGPT only 0.2% > individual users are still mostly in ChatGPT: Codex is only 16.5% of their output tokens > organizations are using Codex like no one else. Codex went from almost nothing in Aug 2025 to 63.3% of output tokens now > long horizon tasks are taking over. 1h+ human-task prompts went from 35.4% of users in Dec to 70.2% in May > 8h+ human-task prompts went from 2.1% in Dec to 25.6% in May >skills are being adopted fast. usage went from 5.4% in Mar to 26.6% in Jun > OpenAI’s 99th-percentile users are running ~71 hours of agent work per day, they’re managing multiple agents in parallel

English
37
38
1.1K
207.8K
Declutter
Declutter@declutterKE·
@dimoreconqueror @moneyacademyKE if Microstrategy is forced to liquidate by this price drop this could be the beginning of the end. if Microstrategy finds money to buy it would be great strategy to buy the dip
English
1
0
1
33
Omondi X
Omondi X@dimoreconqueror·
@moneyacademyKE $58K is the lowest price BTC has gone to in 21 months... I think this is the bottom.
English
2
0
1
278
Moe
Moe@moneyacademyKE·
Bitcoin now falls below $60,000. The drop has pushed its monthly loss to nearly 20%.
English
32
83
746
28.1K
Ahmad Awais
Ahmad Awais@MrAhmadAwais·
＿＿＿＿＿＿ | We’re hiring | |＿＿＿＿＿＿| \ [• •] / —— | | |_ |_ Open roles in my bio, check my profile.
English
26
7
147
12.7K
Declutter
Declutter@declutterKE·
@CommandCodeAI early tests shows its needlessly expensive and only perfoms at GLM 5.2 level
Declutter tweet media
English
0
1
3
247
Command Code
Command Code@CommandCodeAI·
Fugu Ultra is live in Command Code. A multi-agent model from Sakana AI built to orchestrate top models for hard, multi-step tasks. 1M context. Vision. Tool use. Reasoning. Try it out now with /model. *Available in Max plan and above.
Command Code tweet media
English
14
7
118
13.2K
Moe
Moe@moneyacademyKE·
ZXX
14
76
476
23.2K
Yum⋆₊˚
Yum⋆₊˚@yuhasbeentaken·
glm-5.2 is the best-performing coding model a new test compared 5 coding agents across nearly 1,000 real coding scenarios... glm-5.2 ranked #1 overall with a score of 91.9. it finished ahead of sonnet 4.6, kimi k2.7-code, and qwen 3.7 plus... while also costing slightly less per task than sonnet. the only model that came truly close was minimax m3 at 91.4. and minimax was around 30% cheaper than sonnet. open-source coding models are not just catching up anymore... in this test, they took the top 2 spots!!!
Yum⋆₊˚ tweet media
Yum⋆₊˚@yuhasbeentaken

vercel CEO was "almost shocked" by GLM-5.2, it's impressive and made me think if it's another deepseek moment we're preparing to change our router model into GLM 5.2

English
5
2
27
6.2K
Naymur Rahman
Naymur Rahman@naymur_dev·
Testing which open source model is best at design Asked each one to build a landing page with macOS taste using @CommandCodeAI /design →Kimi k2.7 code [left] →GLM 5.2 [middle] →DeepSeek-v4-flash [right] GLM 5.2 is the clear winner. The other two are decent but not close
English
17
5
211
23.4K
Yum⋆₊˚
Yum⋆₊˚@yuhasbeentaken·
running glm-5.2 locally right now is basically madness... the full model takes around 1.51tb of disk space. even the heavily compressed versions require roughly this much total system memory: • 1-bit: around 223gb • 2-bit: around 245gb • 4-bit: roughly 372–475gb • 8-bit: around 810gb that means ram + vram, or unified memory on machines like a mac studio. the recommended inference settings are: • temperature: 1.0 • top_p: 0.95 • thinking enabled by default so yes... glm-5.2 can technically run locally. but for almost everyone, buying a machine with hundreds of gigabytes of memory makes no practical sense yet. better quantization, more efficient inference, smaller models, and cheaper high-memory hardware could eventually change that... because running a model this capable privately on normal hardware would be a massive deal!!!
Yum⋆₊˚ tweet media
Yum⋆₊˚@yuhasbeentaken

glm-5.2 is the best-performing coding model a new test compared 5 coding agents across nearly 1,000 real coding scenarios... glm-5.2 ranked #1 overall with a score of 91.9. it finished ahead of sonnet 4.6, kimi k2.7-code, and qwen 3.7 plus... while also costing slightly less per task than sonnet. the only model that came truly close was minimax m3 at 91.4. and minimax was around 30% cheaper than sonnet. open-source coding models are not just catching up anymore... in this test, they took the top 2 spots!!!

English
3
0
21
2K
Declutter
Declutter@declutterKE·
@CommandCodeAI Exapnd the moat. Let's see what this thing is really capable of.
English
0
0
0
81
Command Code
Command Code@CommandCodeAI·
We’re working on GUI, SDK, ACP, BYOK, Sandboxes, a secret project, and a new deal. So much shipping in June. Y’all ready for Command Code. Send in your requests. What should we prioritize. Something we missed? Command Code with your taste and the best open models!!
English
47
4
173
201.5K

Keşfet

@MrAhmadAwais @samhogan @CommandCodeAI @vipulgupta2048 @Hesamation @dimoreconqueror @moneyacademyKE @elonmusk