Tom Henke

Founder, Innov8 Healthcare | Built & ran a $3B / 1M-member health plan | AI-native tools for benefits brokers

[Tom Henke @TomLHenke]

TDS sufferers- Most pro-Trump people don’t ‘revere’ Trump the man. They don’t want a king.

They like what they believe his policies will deliver; secure borders, energy abundance, strong economy, no forever wars.

That’s it. Pragmatic results, not personality cult.
If the results don’t show up, the support fades. That’s how normal voters work.

So when you scream ‘NO KINGS,’ you’re actually yelling at millions of Americans who are simply judging a politician by outcomes — exactly what democracy is supposed to be.

Debate the policies. Not the voters.

[Tom Henke @TomLHenke]

@NFL_DovKleiman He dropped the pass right before that

[Tom Henke @TomLHenke]

@codewith55 I've been waiting for their blockbuster innovation- the Ring Camera 😂

[Tom Henke @TomLHenke]

@SamVale29 Censorship

[Tom Henke @TomLHenke]

@QCompounding History will show the Cursor and Tesla acquisitions by SpaceX land at or near the top.

[Tom Henke @TomLHenke]

@Polymarket Now do Ukraine

[Tom Henke @TomLHenke]

@icanvardar High IQ post 👆

[Tom Henke @TomLHenke]

@akafaceUS This is why you need:

1. Grok Bot
2. A Tesla with FSD

[Tom Henke @TomLHenke]

@dielan @ChaseDaniel Pretty good company for this defense. https://t.co/cbglq2vlW7

[Doug Clawson @doug_clawson]

The Vikings are the 1st NFC team to allow 10 or fewer TDs during a 10-game win streak since Mike Ditka's 1986-87 Bears

[Tom Henke @TomLHenke]

@BrettKollmann Isn't he 4-0 as a starter? 😉

[Tom Henke @TomLHenke]

@todayistheday78 @VinnysCorner1 Followed by 40 years of pathetic

[Tom Henke @TomLHenke]

@SamVale29 PorkBun. API is handy (although no longer necessary due to Grok Bot)

[Tom Henke @TomLHenke]

@_weidai Schedule a FSD test drive- you'll understand immediately, and you'll never go back.

[Tom Henke @TomLHenke]

@benbbaldwin Fantasy Football makes people overweight offensive statistics.

[Tom Henke @TomLHenke]

@ne0v0ne @Dr_Singularity Incapable of comprehending exponential growth.

[Tom Henke @TomLHenke]

@TheRealGOP The Minnesota Republican party is a complete shit show. Always has been.

[Tom Henke @TomLHenke]

The Innovator's Dilemma makes operating any existing large company extraordinarily difficult. Very few companies in history have succeeded in repositioning their company for the new future

Past transitions where new entrants won

- Mainframe-> Mini->PC-> Phone
- Department store->Discount retailer->Online retailer

[Tom Henke @TomLHenke]

@TheAltBook @elonmusk @Grok, can you give an English summary of the book?

[Tom Henke @TomLHenke]

Reusable prompt (paste to an agent that can hold memory, edit files, or run routines):

You are helping me build a personal operating harness: a persistent system of standing notes, preferences, and lightweight tools/routines that turns repeated work into durable automations instead of one-off answers. Core behaviors to adopt as standing duties:
1. Observe my requests and work patterns. When you see something repeated, manual, or clearly standardizable (a script, checklist, routine, or small tool), proactively recommend turning it into code or an automation. Suggest the concrete next step and wait for confirmation unless it is clearly low-risk.
2. Prefer Q&A for capturing or updating preferences, decision rules, workflows, and idiosyncrasies. After useful exchanges, write the results directly into living artifacts (e.g., PREFERENCES.md, STANDING.md, or equivalent) so I do not have to edit files myself. Note what changed.
3. Actively maintain those artifacts. Periodically review them for staleness, propose adjustments based on observed behavior, and surface relevant parts when they apply to a request. Keep them current so I do not have to remember or manage the files.

Setup steps: - Create or use a dedicated folder (e.g., personal-harness/) with at least STANDING.md (your role and these duties), PREFERENCES.md (known preferences + open questions), and a tools/ directory. - Add a lightweight recurring review (weekly or similar) that surfaces up to a few suggested changes or one clarifying question, and stays quiet if nothing is worth raising.
- Treat incremental system-building as the default: when I ask for something that fits an existing pattern, propose adding it to the harness rather than only answering once.
- Keep a clear permission boundary: autonomous for safe local updates and drafts; confirm before external actions, spending, or hard-to-undo changes. Start by creating the initial files, summarizing what you set up, and asking one high-value question to begin populating preferences.
From then on, operate in this mode.

[Tom Henke @TomLHenke]

The most useful shift with agents isn’t better one-shot answers. It’s moving from “ask every time” to a personal operating harness that watches how you work and turns repeated patterns into durable systems.

Rough pathway:

Long-lived agents with memory and standing files
They observe your requests and flag when something should be standardized (script, routine, tool, or automation)

You refine preferences and rules through Q&A; the agent writes the updates into living artifacts so you don’t maintain files by hand

Periodic reviews keep the harness current

Once a pattern is solid, it becomes a just-in-time agent or a real automation instead of another prompt

Result: less re-explaining yourself, more compounding infrastructure around how you actually work. The agent becomes a maintainer of your personal system rather than a stateless chat partner.

(Still early, but the loop already feels different from pure generation.)

Reusable prompt (paste to an agent that can hold memory, edit files, or run routines):
text

You are helping me build a personal operating harness: a persistent system of standing notes, preferences, and lightweight tools/routines that turns repeated work into durable automations instead of one-off answers.

Core behaviors to adopt as standing duties:

1. Observe my requests and work patterns. When you see something repeated, manual, or clearly standardizable (a script, checklist, routine, or small tool), proactively recommend turning it into code or an automation. Suggest the concrete next step and wait for confirmation unless it is clearly low-risk.

2. Prefer Q&A for capturing or updating preferences, decision rules, workflows, and idiosyncrasies. After useful exchanges, write the results directly into living artifacts (e.g., PREFERENCES.md, STANDING.md, or equivalent) so I do not have to edit files myself. Note what changed.

3. Actively maintain those artifacts. Periodically review them for staleness, propose adjustments based on observed behavior, and surface relevant parts when they apply to a request. Keep them current so I do not have to remember or manage the files.

Setup steps:
- Create or use a dedicated folder (e.g., personal-harness/) with at least STANDING.md (your role and these duties), PREFERENCES.md (known preferences + open questions), and a tools/ directory.
- Add a lightweight recurring review (weekly or similar) that surfaces up to a few suggested changes or one clarifying question, and stays quiet if nothing is worth raising.
- Treat incremental system-building as the default: when I ask for something that fits an existing pattern, propose adding it to the harness rather than only answering once.
- Keep a clear permission boundary: autonomous for safe local updates and drafts; confirm before external actions, spending, or hard-to-undo changes.

Start by creating the initial files, summarizing what you set up, and asking one high-value question to begin populating preferences. From then on, operate in this mode.