andrewacove

Author here. Anthropic shipped Artifacts in Claude Code this week — agent-built HTML pages you share with your team. Surface is a small, open take on a similar idea, but two-way: the agent stands up the page, watches it, and reacts to what you do on it (click, type, submit) without you copy-pasting back to the chat. It installs as a Claude Code / Codex plugin. I wrote up how it compares to Artifacts here: https://aac.social/posts/enhancing-the-effectiveness-of-html... — happy to answer questions.

It's amusing to think that the endgame is that the humans in the loop are parents with credit cards.

I suppose you could never be sure that an agent would explicitly follow your instruction "Don't spend more than $5".

But maybe one could build a tool that provides payment credentials, and you get to move further up the chain. E.g., what if an MCP tool could spin up virtual credit cards with spending caps, and then the agent could create accounts and provide payment details that it received from the tool?

Do you think this'll eventually be trained into the models the way that chain-of-thought has been?

As a reader of Simon's work, I can speculate an answer here.

All "designing agentic loops" is context engineering, but not all context engineering is designing agentic loops. He's specifically talking about instructing the model to run and iterate against an evaluation step. Sure, that instruction will end up in the context, but he's describing creating a context for a specific behavior that allows an agent to be more effective working on its own.

Of course, it'll be interesting to see if future models are taught to create their own agentic loops with evaluation steps/tests, much as models were taught to do their own chain of thought.

Maybe this is their long term plan for getting rid of the camera bump.

It's very difficult to get debt financing if you're running out of cash, and this article is focused on companies that are (at risk of) running out of cash.

Much of this story was published in Proof of Conspiracy in 2019, curated from major media reporting. The media just didn't cover the release at the time.

This can be difficult because the money is often funneled through other entities (as LPs to VC and growth funds, etc.)

I've had more than one VC tell me they couldn't guarantee that there was no Saudi money in their fund, even though they didn't have any direct KSA LPs.

I've been living in vd for the last few months for data analysis. It's been beyond helpful, including for examining ugly nested JSON data. It's been extremely valuable for feature exploration in my dataset.

I loved Whitefish when I was there, and would have considered it for the long term but was concerned about reports of a growing alt-right contingent in town. Have you encountered that?

Even if you have a terrible experience in your doctoral program, much of its outward value is in the program's reputation. Giving a program a negative review hurts the perceived value of your own education.

This sounds perfect for my burning man camp. Now we won't have to sit at the pepper mill debating how to split things fairly.

Your response is flawed. Not wearing a mask is not an essential liberty.

And yes, if large numbers of Americans don't wear masks, that will kill more people than would die if we all wore our masks. It's not theater.

It's also reminiscent of the installation in Seattle, A Sound Garden, after which the band is named. https://www.youtube.com/watch?v=H4Z2fqxkdvk

I lived in New York for a good chunk of Foursquare's life, and while it was magical, at least half of the population hated the concept, for good reason. Every woman I knew who had used it was terrified by the privacy and safety risks engendered by making their location public. The ones who did check in somewhere would typically only do it as or after they'd left. The "magical power" to see through walls that Dennis touted was a nightmare for women, who already deal with constant harassment and stalkers in the city. Which was a shame, because it really could feel magical to meet up with friends because of where you'd checked in.

For a while, Foursquare also had the most reliable review data – or at least its users tastes better aligned with mine. It was refreshing to get useful content about a restaurant without reading through paragraphs of self-indulgent drivel on Yelp. But after the launch of Swarm, the reviews more or less rotted over time.

The founders of Foursquare got to take a lot of money off the table in a relatively early round, ($8M each? Series C?), which is something I'd love to get the present perspective on from investors and employees who were there along the way.

I've had this debate with my academic advisor at CMU.

I'd take a trained speaker–even an actor who doesn't actually know the material he/she's describing–as a primary lecturer any day, as long as I get access to someone who knows what they're talking about in a recitation.

In some cases, yes. I'm going to run this by some of the team and see where it lands.

Thanks James. This is super useful feedback.

I'm still working on how to present the credibility of the team. I've avoided listing everybody on the site thus far because people's availability varies.

I can absolutely get a quote from a VC (or two...). That's a great suggestion. I'm also in the process of adding logos to the site of funds who've invested in companies that we've helped.

I've been thinking about versions of this ever since I was at AngelList.

A lot of companies don't get a good look from investors because they don't understand investor expectations. "Negative signals" kill pitches. It might be a sign that the system works – but I suspect it's an inefficiency. Not every company has access to the kind of advice that YC and other accelerators provide.

The underlying principle is that the bulk of pitch mistakes are common. So Deck.vc reviewers can have a checklist of the most common recommendations and generate the bulk of the review really efficiently. That's what makes it possible to price this like a product, instead of consulting.

My goal is for Deck.vc to be of value to every founder who goes out to raise. Feedback greatly appreciated.

Giving up the band and getting a real job.

One of the cofounders of ITA Software, Dave Baggett, used to work for Naughty Dog, which is one of the top studios in the world for this type of low-level development. Given Dave's PS1 work, it's no surprise ITA values the same skills.

(Full disclosure: I worked at ND).

I find it interesting to contrast this to the requirements for the YC application video:

Please do not recite a script written beforehand. Just talk spontaneously as you would to a friend. People delivering memorized speeches (or worse still, text read off the screen) usually come off as stupid. Unless you're a good enough actor to fake spontaneity, you lose more in the stilted delivery than you gain from a more polished message.

Footnote 2 seems relevant. I'd guess that most YC application videos are also made of spolia.

The data is two years old, but it looks like you're at the start of your traditional March/April/May writing tear: http://andrewacove.posterous.com/wip-pg-essays

I guess he'd have a good answer for the "Tell us about a time you hacked a (non-computer) system to your advantage" question on the YC application.

Is there any good documentation for setting up red5? My first attempt was a nightmare.

I may not have been clear enough in explaining it, but my blog post isn't a direct response to Micah's, though I use his post for context. Mine is a response to the people who think they are hustlers (by Micah's definition), promote themselves as such, and are in actuality the type of people I describe. Unfortunately, they significantly outnumber hustlers who hew to the intended definition.

The post is also tinged with some additional dismay at my discovery that some of the hustlers held up in the valley as prime examples of the form have proven themselves to be sleazy as well. One of those people is who I refer to in the sentences about lies.

> the summary is that good hustlers are hard to come by

That's true, but I'll add to it. I've met good hustlers - people who _get shit done_ - but they still leverage sleazy tactics. And that's the problem. Out of the possible combinations of sleaze and GSD, I keep encountering {sleaze} and {sleaze, GSD}. What I'm finding hard to come by, and what I'd kill for in a cofounder is the other: {GSD}. And tbh, I think {sleaze, GSD} is more representative of what people view as an ideal hustler. I can say, with certainty, that I don't want that in a partner.

That's definitely true. Yours is more in line with the intent of Micah's original blog post.

Mine is based on my experiences with the people I meet who intend to fill the role. In theory, your definition should be correct - in practice, the scene is full of people who represent mine. They might be making an unfounded claim to the title, but filtering the network to just the real ones is a nightmare.

Upvoted. As I've often heard people say about good UX designers, "I'd also like a unicorn."

Honestly, though, I'm more interested in Hacker/Humanist than Hacker/Hustler. I think a UX person would be a much more valuable partner for me.

I got good insurance in CA for $167 a month as a 26 year old last year. This year, the premiums went up to $240.