mewc
@mewc.bsky.social
Stop building another Dashboard. Turn your business into an engine where every pulse is meaningful and impact-dense. Get pulses into channels, and let systems analyse it knowing all past context in docs and discussions so you can free up your brain for higher level things
February 6, 2026 at 3:00 AM
Stop building another Dashboard. Turn your business into an engine where every pulse is meaningful and impact-dense. Get pulses into channels, and let systems analyse it knowing all past context in docs and discussions so you can free up your brain for higher level things
feel your pulse at the doctors, and for your business. When a high-fidelity chart lands in the Slack channel, with high-context analysis, and where your team is already debating strategy the data becomes the center of gravity for the conversation, rather than a link that gets ignored.
February 6, 2026 at 3:00 AM
feel your pulse at the doctors, and for your business. When a high-fidelity chart lands in the Slack channel, with high-context analysis, and where your team is already debating strategy the data becomes the center of gravity for the conversation, rather than a link that gets ignored.
We’ve been conditioned to treat data like a destination - a library you have to physically visit to get an answer. But in a fast-moving org, if you have to "go" somewhere to know how your business is doing, you’re already behind. You shouldn’t have to hunt for the pulse of your own company.
February 6, 2026 at 3:00 AM
We’ve been conditioned to treat data like a destination - a library you have to physically visit to get an answer. But in a fast-moving org, if you have to "go" somewhere to know how your business is doing, you’re already behind. You shouldn’t have to hunt for the pulse of your own company.
...you have to remember the URL, find the login, bypass 2FA, and set your teams filters, and then all of a sudden you missed a drilldown case - the data engineer is away on holiday, and suddenly its forgotten about. By day three, curiosity loses to the path of least resistance.
February 6, 2026 at 3:00 AM
...you have to remember the URL, find the login, bypass 2FA, and set your teams filters, and then all of a sudden you missed a drilldown case - the data engineer is away on holiday, and suddenly its forgotten about. By day three, curiosity loses to the path of least resistance.
and multiple checkouts just mean i have n stacks running at the same time, so ive gotta deal with many envs, parameters, ngrok/smee bits coming in which is totally tedious and just doesn't work for generally wide and interconnected apps (selfcontained oss things like cc/opencode is fair enough
February 4, 2026 at 3:46 PM
and multiple checkouts just mean i have n stacks running at the same time, so ive gotta deal with many envs, parameters, ngrok/smee bits coming in which is totally tedious and just doesn't work for generally wide and interconnected apps (selfcontained oss things like cc/opencode is fair enough
like i find a few bugs, and a few new ideas here and there and bundle things in.
its much simpler to just roll with it and say "this has this main thing, but also a fair bit of everything else"
and just let the overall changeset reviewer come up with a final summary of all the things going on
its much simpler to just roll with it and say "this has this main thing, but also a fair bit of everything else"
and just let the overall changeset reviewer come up with a final summary of all the things going on
February 4, 2026 at 3:46 PM
like i find a few bugs, and a few new ideas here and there and bundle things in.
its much simpler to just roll with it and say "this has this main thing, but also a fair bit of everything else"
and just let the overall changeset reviewer come up with a final summary of all the things going on
its much simpler to just roll with it and say "this has this main thing, but also a fair bit of everything else"
and just let the overall changeset reviewer come up with a final summary of all the things going on
Your value is in the insight, not the assembly. Impact not Charisma (both together do help). Reclaim your time. Setup your chartcastr.com for your pulses
February 3, 2026 at 5:55 PM
Your value is in the insight, not the assembly. Impact not Charisma (both together do help). Reclaim your time. Setup your chartcastr.com for your pulses
Your business pulses should be set workflow. Define the metrics once. Set the distribution schedule. Spend your time actually reading the pulses, chat to different AIs to get feedback since it's done all the thinking and gathering context anyway.
February 3, 2026 at 5:55 PM
Your business pulses should be set workflow. Define the metrics once. Set the distribution schedule. Spend your time actually reading the pulses, chat to different AIs to get feedback since it's done all the thinking and gathering context anyway.
Thats why I believe in pushing charts & proactive analysis into the edges where you already work, and @chartcastr handles it for you - auto link up sources so data pulses ship into channels with the full visualisation, analysis and no more "Give me access please!"
February 2, 2026 at 3:22 PM
Thats why I believe in pushing charts & proactive analysis into the edges where you already work, and @chartcastr handles it for you - auto link up sources so data pulses ship into channels with the full visualisation, analysis and no more "Give me access please!"
Transparency isn't just a culture value - it’s an ROI play and full on acceleration thing. So your business is run on data, and access or mysterious URLs that rarely work is how people need to do their job to see whats going on?
February 2, 2026 at 3:22 PM
Transparency isn't just a culture value - it’s an ROI play and full on acceleration thing. So your business is run on data, and access or mysterious URLs that rarely work is how people need to do their job to see whats going on?
We’ve been gatekeepers for key data for too long. But for an organization to move fast, data needs to be "push," not "pull."
Instead of sending people to a mysterious URL they can't open, push the insights directly into the channel.
Instead of sending people to a mysterious URL they can't open, push the insights directly into the channel.
February 2, 2026 at 3:22 PM
We’ve been gatekeepers for key data for too long. But for an organization to move fast, data needs to be "push," not "pull."
Instead of sending people to a mysterious URL they can't open, push the insights directly into the channel.
Instead of sending people to a mysterious URL they can't open, push the insights directly into the channel.
xAi is a pretty extreme case, so how can we take ultra high performing ideas there and apply it across all other business? how can we build tools to take these ideas and systemise things? no docs, 15 minute not day/week/quarter cycles, experiments, always in person, always shipping
January 20, 2026 at 3:20 PM
xAi is a pretty extreme case, so how can we take ultra high performing ideas there and apply it across all other business? how can we build tools to take these ideas and systemise things? no docs, 15 minute not day/week/quarter cycles, experiments, always in person, always shipping
Just like rev x employee ratios, theres a high per-engineer impact: ~$2.5M per main-repo commit.
January 20, 2026 at 3:20 PM
Just like rev x employee ratios, theres a high per-engineer impact: ~$2.5M per main-repo commit.
Elon? jumps on blockers, calls vendors directly, fixes in hours.Requires experiments (show dont tell), not opinions. Ends meetings with "How can I help?"
January 20, 2026 at 3:20 PM
Elon? jumps on blockers, calls vendors directly, fixes in hours.Requires experiments (show dont tell), not opinions. Ends meetings with "How can I help?"
Daily (or multiple daily) iterations, even from pre-training. "Delete-first" methodology and add back only what's essential. Smaller/faster models prioritized. Still a hard physics-level problem. Heavy postmortems on Grok mistakes.
January 20, 2026 at 3:20 PM
Daily (or multiple daily) iterations, even from pre-training. "Delete-first" methodology and add back only what's essential. Smaller/faster models prioritized. Still a hard physics-level problem. Heavy postmortems on Grok mistakes.
Scaling to 1M+ emulators? Bootstrap off Tesla fleet. 4M+ HW4-capable vehicles idle 70–80% of time, with networking/cooling/power built-in. Owners lease out compute → way more capital-efficient than building datacenters.
Idling Teslas as future AI compute army. Unstoppable
Idling Teslas as future AI compute army. Unstoppable
January 20, 2026 at 3:20 PM
Scaling to 1M+ emulators? Bootstrap off Tesla fleet. 4M+ HW4-capable vehicles idle 70–80% of time, with networking/cooling/power built-in. Owners lease out compute → way more capital-efficient than building datacenters.
Idling Teslas as future AI compute army. Unstoppable
Idling Teslas as future AI compute army. Unstoppable
MacroHard = human emulator. Digital agents that automate any keyboard/mouse/screen task 8× faster than humans, 24/7, no software changes needed. Uses small, fast models for insane iteration speed. Internally already testing with hilarious stories of invites to non-existent desks.
January 20, 2026 at 3:20 PM
MacroHard = human emulator. Digital agents that automate any keyboard/mouse/screen task 8× faster than humans, 24/7, no software changes needed. Uses small, fast models for insane iteration speed. Internally already testing with hilarious stories of invites to non-existent desks.
Colossus supercomputer built in 122 days using a "carnival permit" loophole. 80+ mobile generators, multi-layer power (capacitors/batteries/generators) for millisecond MW failover. Racks start training hours after setup. Elon personally calls Nvidia for next-day patches.
January 20, 2026 at 3:20 PM
Colossus supercomputer built in 122 days using a "carnival permit" loophole. 80+ mobile generators, multi-layer power (capacitors/batteries/generators) for millisecond MW failover. Racks start training hours after setup. Elon personally calls Nvidia for next-day patches.
Culture = extreme talent density + collaborative chaos. One building, fuzzy teams, war rooms with sleeping pods & bunk beds for all-nighters.
January 20, 2026 at 3:20 PM
Culture = extreme talent density + collaborative chaos. One building, fuzzy teams, war rooms with sleeping pods & bunk beds for all-nighters.
No docs (things change too fast; they're experimenting with Grok auto-generating them). Even the sales team? Engineers who train models.
January 20, 2026 at 3:20 PM
No docs (things change too fast; they're experimenting with Grok auto-generating them). Even the sales team? Engineers who train models.
Everything is needed yesterday - No due dates. Aim for 1y projects in 1 month, and often finish in 2. Ruthlessly audit, and only real physics is a blocker. Flat structure: ICs → co-founders → Elon. Anyone can fix anything & merge instantly - live and die by the sword
January 20, 2026 at 3:20 PM
Everything is needed yesterday - No due dates. Aim for 1y projects in 1 month, and often finish in 2. Ruthlessly audit, and only real physics is a blocker. Flat structure: ICs → co-founders → Elon. Anyone can fix anything & merge instantly - live and die by the sword