Sherveen

In this thread: people who would blame the inventors of the computer for the impact of computer viruses.

What an insightful contribution.

Ed Zitron is the most objectively and confidently wrong human re: anything going on in AI, competing only with the likes of Gary Marcus and, on his bad days, Yann LeCun.

People keep asking for this w/ Codex, too, and I really regret that both labs seem inclined to listen.

If you ask a thinking/research type question in 'Chat' versus 'Work' mode in these products -- say, something complex about politics, or do a multiturn business strategy, or want to work thru a new concept, you get very different answers.

The harness, steering, etc. in the chat/reasoning products is so much better for this type of question (that doesn't require code as a primary substrate).

Even as someone who mainlines like, 7 coding agents at all times, I regret that productivity fever will mean the regression of think-first-act-later AI UX.

I have a repo/project called Loadouts & Summons. It has a primary skill, `capsule`.

All skills, MCPs, CLIs, etc. live inside of it. I have it symlinked to all my dev machines so that it doesn't have to be an MCP.

`capsule` is then progressive to dozens of skills/tools thru `capsule` -- ex. `$capsule plannotator [args]`.

In some harnesses, I make it human-invoke only, and call it directly. In others, I let the model invoke it, and it has a top-level description that hints at what's inside.

Maximal context/session start control and capability extension.

Founder and investor FOMO has reached levels of desperation and moral emptiness never seen before. It's so pervasive and powerful that it's unironically responsible for pretty much everything bad in the US over the past 5 years.

I don't understand this critique. (1) Did you previously think you weren't getting paid for doing what a company wants you to do, aka what THEY thought was productive? (2) Do you think all this AI generated code is useless?

Edit: y'all are some whiney folk, ain't ya?

Check out the strats from the Every team: https://github.com/EveryInc/compound-engineering-plugin (I recommend just learning from their builds and doing your own!)

Simon Willison's blog: https://simonwillison.net/guides/agentic-engineering-pattern...

The Every team: https://github.com/EveryInc/compound-engineering-plugin (I recommend just learning from their strats and doing your own!)

Simon Willison's blog: https://simonwillison.net/guides/agentic-engineering-pattern...

As I said on Product Hunt (which upset Garry quite a lot) --

If he weren't the CEO of YC, this wouldn't be on PH, and it wouldn't be on HN.

This is not an impressive setup, folks. It's overengineered and deeply into its own form -- it will not make your agents better, and is likely to make it worse. There are lots of other people to follow/learn from/mimic for skills/context engineering.

I think skills are probably a net positive for the general population, but for power users, I do recommend moving one meta layer up --

Whenever there's an agent best practice (skill) or 'pre-prompt' that you want to use all the time, turn it into a text expansion snippet so that it works no matter where you are.

As an example, I have a design 'pre-prompt' that dictates a bunch of steering for agents re: how to pick style components, typography, layout, etc. It's a few paragraphs long and I always send it alongside requests for design implementation to get way-better-than-average output.

I could turn it into a skill, but then I'd have to make sure whatever I'm using supported skills -- and install it every time or in a way that was universally seen on my system (no, symlinking doesn't really solve this).

So I use AutoHotkey (you might use Raycast, Espanso, etc) to config that every time I type '/dsn', it auto-expands into my pre-prompt snippet.

Now, no matter whether I'm using an agent on the web/cloud, in my terminal window, or in an IDE, I've memorized my most important 'pre-prompts' and they're a few seconds away.

It's anti-fragile steering by design. Call it universal skill injection.

Quote from article:

"The Trump administration announced on Monday that all foreign-made drones and their components posed 'unacceptable risks to the national security of the United States' and would be put on a federal blacklist of equipment makers prohibited from selling their goods freely in the country."

Related FCC fact sheet: https://docs.fcc.gov/public/attachments/DOC-416839A1.pdf

Yeah, everyone else in the comments so far is acting emotionally, but --

As a fan and DAU of both OpenAI and the NYT, this is just a weird discovery demand and there should be another pathway for these two to move fwd in this case (NYT to get some semblance of understanding, OAI protecting end-user privacy).

You're 110% doing something (or many things) wrong.

Don't you think it's a little circular that you always default to assuming that their support is about regulatory capture?

Like, what if they had that opinion before they built the company? If you saw evidence of that (as is the case with Anthropic), would that convince you to reconsider your judgement? Surely, you think... some people support regulatory frameworks, some amount of the time... and unless they banned themselves from every related industry, those might be regulatory frameworks that they might one day become subject to?

You aren't taking what he said seriously. The junior could also get sick, present management issues, etc.

If this person plus a junior represented "1.3 engineering knots," he's saying... "actually, I'm still 1.3 engineering knots without him."

When this person leaves, they go find someone else who is 1.3 engineering knots. The junior represented .3, without the 1., it doesn't matter that much. Headcount strategy shifts.

... and not everyone will seek to be the best at prompting.

This is so wrong... there is so much surface area for people to be better at flex-prompting consumer- and enterprise-grade applications.

Love it. I've got a lot of these sorts of micro-tools, too, CLIs, etc.

Even better when you have them all in a repo w/ an agent like Codex or Claude Code to constantly tweak/remix them as needed.

This makes a tremendous amount of sense. Most people are so bad at using AI for productive purposes -- but outside of eng, most AI fluency is actually hot garbage. People just haven't gained an understanding or appreciation for the degree of quality and capability they can achieve.

And once they can use it and get visible results, those orgs are ripe for large amounts of AI product adoption.

Only downside to the OAI version will be that it's OAI specific.

Neat.

As someone who has tried almost all of the AI browsers that are accessible or in a relatively open beta, plus all the browser control frameworks and agents, I super agree with the notions behind this post.

Curious about your approach, though: so, it's a literal script, or an LLM being told to follow a deterministic script and only get subjective when necessary? Based on the blog, it looks like the former, but why not the later? Get the LLM to be pseudo-deterministic but still step-by-step it so that it can handle UI changes and adjacent interfaces.

Ugly. Just like their new logo and their president.

Everyone in this thread who posts some variation of "wow love it how the government gets to decide if you get to sell your startup or how the market should work" should be handcuffed to their chair and forced to answer these 3 questions:

1. is there any role for gov't antitrust in your view of modern capitalism? 2. if there is a role, why is Adobe x Figma not the perfect example for enforcement? 3. if your answer is "Adobe clearly isn't a monopoly, look at the existence of Figma as evidence," why are you dumb?