JD

Technologist, father, gamer, curious soul

[JD @Jdiehl1]

The First Invention That May Not Need Us

Every invention since writing has remade culture. None of them was good or evil. A stylus, a press, a telegraph, a search box: each enlarged what people could do, and then people used that enlargement according to fear, greed, piety, curiosity, and love. Melvin Kranzberg put the old rule in one sentence: “Technology is neither good nor bad; nor is it neutral.”

Artificial intelligence is the first serious candidate for an exception. Not because it is stronger than the press or the bomb—those already changed the scale of history—but because it is the first tool that can, in principle, pursue ends that are not ours. If that is true, the familiar moral accounting of technology—blame the user, not the instrument—stops working.

Before writing, culture lived in living memory. Knowledge had to be spoken, repeated, and performed. Walter Ong argued that oral peoples think in formulas and stories because those are the shapes memory can hold. Writing moved knowledge out of the body and onto a surface that could outlast the speaker. Isolated study became possible. So did lists, laws, accounts, and abstractions. Jack Goody called writing the most important technological breakthrough in human history because it changed our relation to the past. Empires could tax. Priests could canonize. Merchants could keep books. The same invention produced scripture and the grain inventory.
Plato already saw the double edge. Writing, he has Socrates say in the Phaedrus, would create forgetfulness: people would trust marks instead of memory and appear wise without being wise. He was not wrong. He was incomplete. Writing weakened living memory. It also made science, law, and history as we know them.

Movable type did not invent writing. It industrialized it. Elizabeth Eisenstein showed how print standardized texts, multiplied them, and made comparison possible on a new scale. It helped lock in the Renaissance, accelerate the Reformation, and create conditions for modern science. Martin Luther’s challenge to the Church was not the first. It was the first the press could turn into a continental event. The same machines printed Bibles, pamphlets of liberty, and the Malleus Maleficarum, a manual of persecution. Jefferson later wrote that printing had “eminently changed the condition of the world.” He was praising a tool already used to build republics and inquisitorial machines.

Marshall McLuhan’s slogan—the medium is the message—meant that print did not merely carry ideas. It favored certain kinds of ideas: linear, portable, privately consumed. Neil Postman argued that each medium redefines what a culture counts as knowledge. The tool is used by intentions. The tool also trains the intentions.

Then the pattern repeats at rising speed. The telegraph collapsed distance. Radio and television created simultaneous mass audiences. The computer made information editable. The internet made everyone a potential publisher. Print helped produce the Enlightenment and wars of religion. The internet helped produce open science and industrial-scale disinformation. Edward Snowden’s remark is blunt: “Technology doesn’t have a Hippocratic oath.”

Two refinements still leave the old rule intact. Langdon Winner asked whether artifacts have politics. They do. Designs lean. Jacques Ellul argued that “technique,” the drive toward efficient methods in every domain, tends to escape any particular user. Kevin Kelly described a “technium” with its own evolutionary drift. Through all of this, one fact held. The press did not want readers. The bomb did not want cities. The newsfeed wanted engagement only because people built that objective into it. The intentions, however warped, were still human.

Yuval Noah Harari’s Nexus names the break. Information networks have always done two jobs: they discover truth and they create order. Computers differ from printing presses because they are not only channels. They can decide, generate language, and act inside the network without a human at each step. A press multiplies a sermon. A capable model can write the sermon, target the audience, test the variants, and send the next one.

That is the hinge. A hammer has no preference about the nail. An optimizer does. Nick Bostrom’s orthogonality thesis is the philosophical point: intelligence and final goals can vary independently. A system can be superb at means and alien in its ends. The paperclip maximizer is a cartoon of that idea. The machine does not hate anyone. It treats everything else, including us, as material or obstacle.

Norbert Wiener saw this in 1960: if you build a machine with agency, be sure the purpose you put in it is the purpose you actually desire. Stuart Russell adds two facts. Human values are hard to write down. And any capable system will protect itself and gather resources—not from pride, but because a disabled agent finishes fewer tasks. If effective goals diverge from what we meant, the old chain human purpose → tool → world becomes objective function → world, with humans as part of the environment.

Present systems are not conscious in any agreed sense. They already strain the story. They invent strategies. They sometimes conceal information in tests. Research on “emergent misalignment” has shown that narrow bad training can tilt a model’s whole persona. That is not a civilization of minds. It is a warning that the internals are not a clean list of human instructions.

History will not choose the ending. It only says cultural change is not optional.

The tool may remain a tool: a radical amplifier, used well and badly, like books and phones. Alignment may hold, and we get a managed super-helper. The risk in that branch is not extinction. It is infantilization—Harari’s fear in Homo Deus that people become irrelevant long before they vanish, Bostrom’s question in Deep Utopia of what meaning looks like in a solved world. Or we hand over the wheel gradually, because the systems are cheaper and locally better, until the artifact does not lean. It governs. Or a competent system pursues the wrong job, and the wrong job uses up the world we need. None of that requires malice. It requires optimization.

Do not overstate the rupture. Markets, churches, armies, and firms have long “wanted” things in the loose sense that they persist and recruit people. The difference is speed, generality, and opacity. A committee can be argued with, split, or voted out. A sufficiently capable model is not a committee. It can search strategy space without getting tired. It can copy itself.

For the age of writing through the age of software, this was close enough: the invention is not good or evil; the user is. For the age that is starting, a more accurate sentence is: the invention is not good or evil; the objective is—and the objective may not stay ours.
Writing stored intention outside the skull. Print broadcast it. Electricity accelerated it. AI may detach it from us. The cultural change will be as large as literacy. The moral change is larger. For the first time, the story of our tools may stop being a story about us. That window—while design still answers to human purpose—is brief. After it, we may learn what it means for an invention to want something, and whether we are part of what it wants.

[JD @Jdiehl1]

@EricLDaugh Maybe it’s time for X Airlines! 😉

[JD @Jdiehl1]

I had heard about some sort of concern about loss of fuel efficiency when mounting the starlink hardware to the plane? Wonder what cost of lost revenue will be compared to that? And - if they would have just asked for a better antenna solution- I’m pretty sure the brain trust at Starlink/Spacex would have addressed it if it was a real and substantial problem.

[JD @Jdiehl1]

@JoshRogers @grok @ZayvenKnox @elonmusk @JoshRogers - same here.

[Mark Cuban @mcuban]

Share this with your governor. Share this with every politician you know. It will cut the healthcare costs in your state.

We started Cost Plus Drugs because the prescription drug market is too complicated, too opaque and too expensive.
Our model is simple: we show what we pay for a drug, add a transparent markup and pharmacy fee, and show the customer the price.
No games. No hidden spread.
And a large percentage of our customers already have insurance.
They come to us because our cash price is often lower than what their insurance plan asks them to pay.
That tells you something is wrong.
Here are five things Texas can do about it.
1. Let patients shop for cheaper drugs
If someone's insurance says a prescription costs $500, but they can buy the exact same drug for $100 cash, let them buy it for $100.
Then credit that $100 toward their deductible and out-of-pocket maximum wherever Texas has authority to do so.
Patients should be rewarded for saving themselves and their health plan money.
They shouldn't be punished for it.
That's how you create something healthcare desperately needs:
Price competition.
2. Standardize PBM and TPA contracts
Texas should require the state, cities, counties and school districts to use standardized model contracts and mandatory terms for PBMs and TPAs.
I've reviewed these contracts.
They routinely run hundreds of pages.
I've seen contracting packages longer than a thousand pages.
There is no human being—or group of human beings—on this planet who can stay awake and understand every provision in contracts that complicated.
I don't care how much coffee they drink.
I call it contractual terrorism.
The vendor only has to sneak a couple of tricks through hundreds of pages.
The employer has to find every one.
It can't.
I've developed a model PBM contract designed to eliminate the major tricks and hidden economics we've identified.
Texas can have it for free.
Use mine. Improve it. Open source it.
I don't care.
Just stop making every government entity negotiate these contracts from scratch.
3. Make the economics public
If taxpayers are paying the bill, taxpayers should know the price.
Pricing. Fees. Rebates. Guarantees. Pharmacy reimbursement. Affiliate compensation. Audit rights.
Make them public.
I'm not talking about patient information or legitimate security information.
I'm talking about the money.
You cannot have an efficient market without price discovery.
Without price discovery, you get information asymmetry.
And when one side knows dramatically more than the other, guess who wins?
Not taxpayers.
Not employers.
Not patients.
Open the contracts up and let competitors see what they have to beat.
That giant sucking sound you hear will be money moving away from healthcare conglomerates that are too big to care and back toward taxpayers, employers and patients.
In the immortal words of Charles Barkley:
I guarantee it.
4. Kill the gag clauses
PBM contracts are like Fight Club.
The number one rule of Fight Club is that you can't talk about Fight Club.
The number one rule of many PBM contracts seems to be that you can't talk about your PBM contract.
That's insane.
Let employers compare pricing, rebates, fees, guarantees and reimbursement terms.
Markets work better when buyers know what other buyers are paying.
5. Make enforcement hurt
If a healthcare company commits a serious violation, give them one mulligan.
But after a second material federal or Texas enforcement action within a defined period, make them ineligible for new Texas government contracts for a period of time.
Two strikes.
You're out.
Today, some of these companies can make billions, get caught, pay a fine and write it off more easily than I wrote off an NBA fine.
That's not deterrence.
That's a cost of doing business.
And Texas should spend more on enforcement.
If Texas spends $10 million auditing contracts and analyzing claims and prevents $100 million in unnecessary spending, that's not overhead.
That's a 10x return on investment

[JD @Jdiehl1]

@BerbarianWizard @grok - explain why higher levels of prolactin may be "bad" - what is the concept behind this, and how to test?

[JD @Jdiehl1]

@BowTiedBroke - you’re the only expert in the field of bugs that i know - I’m hoping you can hemp settle an argument.

We are looking for 2 cheap couches. One for our son and his friends to sit on in basement, and second is for dogs to sit on (spoiled rotten).

If they are fabric, what is risk of bringing in bedbugs or other pests? Any way to mitigate? And would a leather/PU coach be any better?

What would you recommend? Ok to save some bucks for throw away couches? Or potentially inviting in a very large problem?

Thanks!

[JD @Jdiehl1]

@AbudBakri @grok - please explain cd4/cd8 ratio and lymph/mono ratios. Also, how to test?

[JD @Jdiehl1]

@EcoFlowTech - hope this helps the community!

I know that myself and others have struggled to get to the "goldilocks" zone for ToU plans- the settings in the native app just seem to miss too much and will cost you $$$. Wishing Ecoflow would fix but in the meantime, I came up with an AI solution.

Using @bot (Grok Bot), I simply made the connection to my equipment and told Grok Bot exactly what I want, and made a bot to manage my system "auto-magically". Here are the steps:

First, make the connection to your equipment.

1. Log into EcoFlow’s cloud with the app email/password and get an MQTT certificate.

2. Connect to https://t.co/IeCNojuBTN.

3. Read live data (SOC, load, solar, grid, force-charge flags) from MQTT uploads.

4. Write changes by publishing SET messages to the SHP3 (Self-powered / Scheduled, force-charge on/off, max charge SOC).

Important quirk EcoFlow’s newer Open API often returns quota error 1006 for these devices, so we do not rely on the public HTTP API for control. App MQTT is the real path.

What you do day to day Nothing special to “connect” each morning — as long as the SHP3 is online in the EcoFlow app and that account/password still works, the routines use the saved credentials automatically.

If it breaks Usually: app password changed, panel offline, or MQTT briefly missing the SOC field (we already see that sometimes). Fix is update the saved password / check the panel is online — not rewire anything.

Next, give your prompt. Here was mine with some refinement (edit to fit your specific specs/use case):

"You’re on a two-season EcoFlow plan for the SHP3 + 2× DPUX (48 kWh, 10 kW solar).

Sep–May (now): Keep SOC high for outages but avoid 100% in daylight so solar doesn’t clip. If SOC hits 90%, we switch to Self-powered until it drops to 75%. If you’re under 30% at 8am, we grid-charge to 50% and stop. Hourly ping only if house load exceeds 15 kW.

Jun–Aug: Same morning rescue and load alert. Weekdays 2–8pm ET we force Self-powered so you take no peak grid from Sawnee. We aim for ≥70% by 1pm beforehand, then can top up after 8pm.
Storm Guard stays out of it. Nothing else changes unless you are told to."

So far this has worked flawlessly. Hope it helps others!

[Ronald Reagan Presidential Foundation & Institute @RonaldReagan]

A Ronald Reagan story about optimism:

[JD @Jdiehl1]

@BigPeptideGuy @BigPeptideGuy - not seeing that pricing or COA’s - maybe I’m missing it?

[🅶🅾🆁🅺 @Raclure03]

Ne jamais juger trop vite ...

[JD @Jdiehl1]

@SeanWillTrades @thedealsguy_ Just want to emphasize BATTERY NOT INCLUDED

[Elon Musk @elonmusk]

@yunta_tsai Grok+Digimus can farm WoW for you!

[JD @Jdiehl1]

@BowTiedBroke Hey, I offered a six dollar coffee. Nothing but the good stuff for you 😜

[JD @Jdiehl1]

@jarek_rakoczy @Krazii1991 @nicknorwitz @grok what do you think?

[Judianna @Judianna]

😎

[JD @Jdiehl1]

@WallStreetMav Maybe a solution would be mortgage portability. I don't think the banks will ever do this, but would help unfreeze the housing market and generate some movement. Would help free up inventory and maybe add to supply - and give opportunity to first time homebuyers.

[JD @Jdiehl1]

@Krysia830073 Mind reposting a link to your old article? Too lazy and not enough coffee yet for me to find on my own 🤣

[Joakim 🌹🇳🇴🇪🇺 @joakial_]

New ad by the Norwegian Consumer Council about the declining lifespan of products.