Hacker News: kenjackson

New comment by kenjackson in "Small models also found the vulnerabilities that Mythos found"

kenjackson — Sun, 12 Apr 2026 15:46:27 +0000

I don’t think these companies are hurting for access to code.

New comment by kenjackson in "Scientists invented a fake disease. AI told people it was real"

kenjackson — Fri, 10 Apr 2026 10:38:48 +0000

You can set your prompt to do that. You can have it be extremely skeptical. You can even make it contrarian, if you wanted to be extreme. My current prompt challenges me often, and wants to find weaknesses in my argument.

New comment by kenjackson in "Scientists invented a fake disease. AI told people it was real"

kenjackson — Fri, 10 Apr 2026 10:36:32 +0000

LLMs will need to develop a notion of trustworthiness. Interesting that part of the process of learning isn’t just learning, but also learning what to learn and how much value to put into data that crosses your path.

New comment by kenjackson in "Netflix Prices Went Up Again – I Bought a DVD Player Instead"

kenjackson — Thu, 09 Apr 2026 20:39:48 +0000

The economics for the company aren't great for people that make high frequency use of it. And I suspect that people that would pay for such a service nowadays would make good use of it.

Regarding the 5d guarantee -- I suspect that most disks would show up in 2-3 days, but if you're going to guarantee you'll need some buffer (as I think US Mail says first class is 1-5 days). And I think Netflix was just counting on it mostly being shorter (and may have even had distribution centers at some point in its history).

New comment by kenjackson in "EFF is leaving X"

kenjackson — Thu, 09 Apr 2026 19:11:12 +0000

Protected speech can be beyond politics. Politics doesn't subsume all protected speech.

New comment by kenjackson in "Project Glasswing: Securing critical software for the AI era"

kenjackson — Tue, 07 Apr 2026 21:24:13 +0000

So private companies shouldn’t get to determine who they provide services to? Assuming no extremely malicious intent, I’d be fine if they said it was only going to McDonalds because the founders like Big Macs.

New comment by kenjackson in "Peptides: where to begin?"

kenjackson — Tue, 07 Apr 2026 13:17:59 +0000

I think it’s also that contrarianism generates an argument they can follow - it’s often much more simplistic along some axis. For example, flat earthers superficially have a really simple model. Throw a ball up, of course it comes down. You look straight ahead and it looks flat. Ask them how GPS works and they can’t follow the math anyways.

New comment by kenjackson in "We found an undocumented bug in the Apollo 11 guidance computer code"

kenjackson — Tue, 07 Apr 2026 12:59:37 +0000

While I agree with the sentiment, using AI to write the final draft of the article isn’t cheating. People may not like it, but it’s more a stylistic preference.

New comment by kenjackson in "The case for zero-error horizons in trustworthy LLMs"

kenjackson — Thu, 02 Apr 2026 20:22:27 +0000

OK, I didn't think to disable switch to thinking (didn't know this was a mode). When I did that then it did get it wrong -- oddly it took about the same amount of time, so thinking mode wasn't taking longer, but it was more accurate.

New comment by kenjackson in "The case for zero-error horizons in trustworthy LLMs"

kenjackson — Thu, 02 Apr 2026 18:30:11 +0000

Weird. I tried in chatGPT auto and it worked perfectly. I tried like 10 variations. I also did the letters in words. Got all of them right.

The one thing I did trip it up on was "Is there the sh sound in the word transportation". It said no. And then realized I asked for "sound" not letters. It then subsequently got the rest of the "sounds-like" tests I did.

Clearly, my ChatGPT is just better than yours.

New comment by kenjackson in "The case for zero-error horizons in trustworthy LLMs"

kenjackson — Thu, 02 Apr 2026 17:41:20 +0000

Whenveer I see these papers and try them, they always work. This paper is two months old, which in LLM years is like 10 years of progress.

It would be interesting to actively track how far long each progressive model gets...

New comment by kenjackson in "I created my first AI-assisted pull request"

kenjackson — Tue, 24 Mar 2026 04:24:35 +0000

I learn a lot from code I read, but don’t write. Did the author not read the code and simply threw it over the fence?

New comment by kenjackson in "Walmart: ChatGPT checkout converted 3x worse than website"

kenjackson — Mon, 23 Mar 2026 17:18:59 +0000

Not revisionist at all.

or Netflix, or Amazon, or Uber, or PayPal, etc… Netflix and Amazon both were competing against brick and morter that were everywhere. Blockbuster was in every town, usually in every major neighborhood. The thought was that on Friday night people wanted to get a movie they wanted, not just happen to have the movie that was shipped to them. And then with streaming it was "the content on Netflix is old and dated, who would want this?" They slowly ate from below. Blockbuster scrambled with their own mailed disc offering. And died before it even had a chance to confront streaming.

Repeat this story with B&N where people said that you had to browse the books physically. You couldn't just blindly order online and wait two weeks to get the book (remember they got big before "2 Day Prime").

With PayPal it was about "they don't understand banking or payment -- and it wants to be both?!".

For this OpenAI experience, it doesn't sound great. I have accounts with these places I buy things from. I want to make sure I get my Prime shipping and digital discount via using the Amazon app. But if you could find a way to integrate my accounts all into ChatGPT things might be different. In the same way I used to never use Apple Wallet, but now it really is my go to place for everything I have a card for. I don't have to worry about having my grocery loyalty card or my football season tickets with me or my car insurance card. It's all in wallet. The Apple Wallet sucked until it was suddenly great.

New comment by kenjackson in "Walmart: ChatGPT checkout converted 3x worse than website"

kenjackson — Mon, 23 Mar 2026 15:14:09 +0000

This is all valid (except probably the last sentence), but it also describes so many attempted changes right until they become darn near the default.

This sounds like why I heard Redfin wouldn’t work, or Netflix, or Amazon, or Uber, or PayPal, etc…. There are always these business complexities that make it seem like these spaces have too much friction, but if there’s enough money - if it can be done then people will figure it out.

New comment by kenjackson in "Our commitment to Windows quality"

kenjackson — Sat, 21 Mar 2026 01:50:28 +0000

I’ve read your posts for the past 25 years - originally on slashdot (not literally you). As you proposed, I think you’re fundamentally wrong. I got my MIL a Chromebook and it was probably the single worst technical support decision I ever made. For some, it will always be the year of Linux on the desktop. But rather the reality is the desktop will run its course before Linux has a foothold there.

New comment by kenjackson in "A sufficiently detailed spec is code"

kenjackson — Thu, 19 Mar 2026 06:16:17 +0000

Code is usually over specified. I recently used AI to build an app for some HS kids. It built what I spec’wd and it was great. Is it what I would’ve coded? Definitely not. In code I have to make a bunch of decisions that I don’t care about. And some of the decisions will seem important to some, but not to others. For example, it built a web page whereas I would’ve built a native app. I didn’t care either way and it doesn’t matter either way. But those sorts of things matter when coding and often don’t matter at all for the goal of the implementation.

New comment by kenjackson in "'The Secret Agent': Exploring a Vibrant, yet Violent Brazil (2025)"

kenjackson — Tue, 17 Mar 2026 19:50:34 +0000

Well I think there are some people that disagree.

Ask HN: Best practices or example workflows for agentic development (March 2026)

kenjackson — Mon, 16 Mar 2026 17:53:04 +0000

I'm trying to get really serious about developing with the help of agents. Ideally, I'd like to be somewhat close to the bleeding edge, but probably not right on the edge. Note, this isn't to build production code.

It seems like the type of area where things have changed so fast that stuff written in December is probably out of date now. What are practices or links to pages that describe practices that I should use?

Comments URL: https://news.ycombinator.com/item?id=47402366

Points: 2

# Comments: 0

New comment by kenjackson in "What happens when US economic data becomes unreliable"

kenjackson — Sat, 14 Mar 2026 17:59:56 +0000

It may be unreliable to you. I see the life of most people around me getting better. Even people that are somewhat poor (not dirt poor, but free lunch poor) have homes, three squares and snacks, PS5, mobile phones with cellular data, and cable tv. The biggest life issues I see are usually strongly related to substance abuse and mental health.

Transplanting to even just the 80s would be a culture shock for most people.

New comment by kenjackson in "Universal vaccine against respiratory infections and allergens"

kenjackson — Wed, 11 Mar 2026 03:34:45 +0000

Allergies are not simply overactive immune response. It’s the wrong type of response. What’s really intriguing is how much we can do innate immunity that we have done relatively little with.