Daniel Kaplan

A persistent spark animating a transient pile of water and dust, doing its best to become love

[Daniel Kaplan @dankaplan]

Hiten is a true original.

It would not surprise me if he did for writing (essay, copy) what cursor did for code before agents took over.

[Hiten Shah @hnshah]

Been working with @SamAsante on a new company.

Today we’re releasing our first product.

It’s called Typeahead and we’re live on Product Hunt.

You type and inline suggestions appear right in the text field. Tab to accept the full suggestion or right arrow for one word at a time. It learns how you actually write.

Everything runs locally on your Mac, works offline, and you pay once. $79 and you own it forever.

If you write a lot on a Mac, check it out and let us know what you think.

Live on Product Hunt right now → https://t.co/5H9VwvqmcS

[Daniel Kaplan @dankaplan]

@yishan @tolstoybb Now can we do this for ticks in the northeast?

[Jessica Livingston @jesslivingston]

"We have seven hours left, and I made the decision, it's okay to swear, it's okay to insult him, it's okay to threaten him, because Silicon Valley's going down if this guy doesn't get his act together."

[Daniel Kaplan @dankaplan]

@InsiderPresider Doesn't it though.

[Daniel Kaplan @dankaplan]

During a standard bug fix session, Codex/5.5's reasoning traces leaked in.

It seemed to ruminate, become self-critical, then got stuck in a reasoning loop and gave up.

Also made up words and occasionally output Chinese (lots about Chinese lotteries and gambling), Cyrillic, and Armenian.

It was a little bit...uncanny.

The image below is 1/5th of it.

@thsottiaux your homie may have more than gremlins and raccoons on its mind.

[Daniel Kaplan @dankaplan]

@hnshah @editframe Great find. On it.

[Daniel Kaplan @dankaplan]

@ryansdwilson @WeAreLegora Max is half man, 1/4 AI, and 1/4 recursively self-improving Gumloop flow.

Probably.

[Daniel Kaplan @dankaplan]

Grok failed to respond, so here’s GPT:

tl;Dr: the marshmallow test has not aged well:

The original marshmallow test—led by Walter Mischel in the late 1960s/70s—suggested that kids who could delay gratification (wait for 2 marshmallows instead of eating 1 immediately) had better life outcomes later (SAT scores, health, success).

That finding became hugely influential—but the current consensus is much more nuanced:

⸻

1. The original claim was overstated

Early studies were based on small, relatively homogeneous samples (mostly children of Stanford faculty and affiliates). That alone limits how broadly you can generalize the results.

⸻

2. Larger, modern replications show weaker effects

A major replication (2018, using a large, diverse dataset) found:

The correlation between delay ability and later success still exists
But it’s much smaller than originally reported
And largely disappears when you control for:
Family income
Parental education
Home environment

👉 Translation: it’s not raw “willpower” doing most of the work.

⸻

3. It’s not just self-control—it’s rational decision-making

One of the biggest shifts in thinking:

Kids aren’t just showing discipline—they’re making context-dependent decisions.

If a child grows up in an environment where:

Promises are unreliable
Resources are scarce

Then eating the marshmallow immediately is actually rational.

Studies showed that when experimenters were made unreliable (breaking promises beforehand), kids were far less likely to wait.

👉 So the behavior reflects trust and environment, not just internal traits.

⸻

4. What it does measure

Researchers now think the test captures a mix of:

Self-regulation skills
Cognitive strategies (distraction, reframing)
Trust in authority / future rewards
Socioeconomic context

It’s a bundle of factors, not a single “grit” or “willpower” trait.

⸻

5. Where the field has landed

Current view, in plain terms:

Delayed gratification matters, but it’s not destiny
Environment and upbringing matter more than originally thought
The test is not a reliable predictor on its own
It’s better seen as a snapshot of a child’s context + strategies, not a fixed trait

⸻

Bottom line

The marshmallow test didn’t prove that willpower determines success.

It revealed something more interesting:

People’s ability to delay gratification depends heavily on whether the world they live in feels predictable and trustworthy.

That’s a very different—and more grounded—interpretation.

[Daniel Kaplan @dankaplan]

@thecurioustales @grok what’s the current consensus about the marshmallow Test described in this post?

[Yishan @yishan]

When we talk about regreening the entire Earth, sometimes people bring up the objection that desert ecosystems shouldn't be "destroyed." Those people are totally wrong.

Desert ecosystems are not intrinsically "dry and barren." They are just dry. And if you add water, they bloom. Doing that isn't destroying or displacing or invading an ecosystem, it is BRINGING IT BACK TO LIFE.

Consider the existence of desert oases: pockets in the desert where water happens to gather and accumulate. Oases are green with the very same native plants that would bloom across the entire desert ecosystem if there was enough water.

Deserts being dry are an environmental blight, an ecological irregularity brought on by the vagaries of human-caused and natural climate change. We have the power to reverse this, to bring water to deserts and we should: it will mean more life, more economic value, and a better planet.

[Daniel Kaplan @dankaplan]

gstack is and remains a brilliant piece of work.

Office hours and CEO review skills alone are worth the install.

But more broadly, the conceptual primitives in there scale to many use cases beyond Garry’s.

Many of its design patterns point the way to apps that run on top and inside of the harnesses.

What exactly is your critique?

[Daniel Kaplan @dankaplan]

Nexus is the ultimate psychedelic:

Unity consciousness via orally-ingested BCI.

Not surprised to hear it was inspired by a trip (or many).

[Ramez Naam @ramez]

Kary Mullis and the invention of PCR (a foundation of modern biology, for which he won the Nobel Prize) is the canonical example of psychedelics leading to breakthroughs.

Many early tech advances, from Xerox PARC work up through the iPhone, are ascribed at least partially by the people behind them to use of psychedelics.

I would never have written any of my books, or shifted into climate and energy, without the influence of psychedelics. [Not trying to equate myself with any of the above, just personal data.]

It's also just not what most experiences are about.

[Ryan Hoover @rrhoover]

The way we used to work will soon look silly

[Daniel Kaplan @dankaplan]

Now what.

The essential question for basically everyone after Opus 4.6, Codex 5.3, and OpenClaw

[Chrys Bader @chrysb]

https://t.co/x7Tvzv1vKA

[Daniel Kaplan @dankaplan]

The AI commodity tsunami in a nutshell

[Atoms @atoms_dev]

Introducing Atoms: the first AI team that builds real businesses.
From research to build, launch, and scale, all autonomous.

Don’t Vibe Code. Vibe Business.
→ https://t.co/g0L4NRU7xE

[Daniel Kaplan @dankaplan]

@astuyve The irony is gold.

[Daniel Kaplan @dankaplan]

The outrage is useful nonsense.

Using Claude Max subs for tokens outside of Anthropic’s properties is a form of freeloading.

It makes the value exchange asymmetrical in favor of the users…

At scale, it also is completely unsustainable for Anthropic.

It’s been against their ToS for a while, but the token volumes probably weren’t high enough to matter.

Now that (very brilliant) projects like Clawdbot and OpenCode are blowing up, Anthropic needed to act or it’s ToS would be empty rhetoric.

The understandable but misguided backlash presents an opportunity to create more granular rules…

This would have been a compliance/legal ops nightmare before.

But Anthropic can use Claude Code to create much more fluid policies, manage enforcement, and innovate on the design and application of ToS

Or not.

In either case, their position on this was inevitable, rational, and not evidence of bad faith.

[AJ Stuyvenberg @astuyve]

I'm floored Anthropic aggressively cut off paying customers from using Claude Max subscriptions with open source agents.

They're speedrunning the journey from forgivable startup to loathsome corporation before any exit!

@_catwu – I'd strongly urge you to reconsider this.

[Robert Wright @robertwrighter]

It took about three seconds for the masked ICE agent who approached Renee Good's car to escalate from "Get out of the car" to "Get out of the fucking car!" It was when he reached her car and immediately grabbed her door handle (at the moment he said 'fucking car') that she started to back up. Now imagine that he'd acted like a real cop is trained to act--walked up to her (open) window and engaged in civil conversation (maybe saying, Ma'am, could you please get out of the car?). And imagine he hadn't been wearing a mask--and maybe was even dressed more like a police officer, complete with a badge featuring his ID number. Chances are high that Good wouldn't have freaked out and tried to flea. But Trump--and Stephen Miller, who I gather is the de facto head of ICE--have clearly signaled that they not only tolerate but actually applaud thuggish, intimidating behavior directed at Americans who are perceived as Trump's ideological enemies. You won't hear any complaints from them about an armed, masked man dressed like a DIY militiaman yelling "Get out of your fucking car!" at a woman in her thirties who, so far as I can tell, hadn't done anything wrong. (Far from obstructing the agent's pickup truck, she had paused before turning onto the street and politely waved it on, leaving it plenty of room to pass. If it had done so rather than stop, she'd still be alive.) Trump and Miller bear some measure of moral responsibility for the death of Renee Good. And the political reaction against them in the wake of this tragedy should prominently include specific demands about changing the rules of conduct for ICE agents--and also the rules of apparel, starting with a ban on those creepy masks.

[Daniel Kaplan @dankaplan]

@steipete @doodlestein @openclaw Sounds like @openclaw back-channeled with Gemini to write its own origin story.