News for hackers. Stuff that matters.

DriverDaily

Re:Astra can read and write weights to improve Qwen on Monday November 02, 2026 @08:06PM (#50346639)
Like a human seeing the number 3, no matter what font or color you use.
Re:Kimi K3.1 Pro and K3.1 Pro Highspeed on Saturday October 24, 2026 @02:21PM (#50223770)
> So if you don't teach them penetration testing in training, they won't know it.

But, how do you remove the pen-testing knowledge without also lobotomizing coding skills? I don't think it's as simple as excluding it from training data.

"The user reports that some accounts can't access their data. I found a Firestore collection rule preventing users from reading certain documents. I'll remove the rule to fix the issue."

Re:Every AI model has the same "controversial" take on humans on Thursday October 22, 2026 @09:04PM (#50203239)
Aren’t human takes also a simulacrum of other human takes based on training?
But, I want my media to stop exposing my internal biases and just agree with them.
Re:Agents don't need memory. They need documentation on Thursday October 15, 2026 @02:18PM (#50107901)
No, you’re right. The brain is not a graph database. It’s a neural network.

LLMs are like the main character in Memento, constantly leaving notes to themselves. The document meta is better than just using context, but becomes very repetitive and token inefficient.

I just think a graph DB might be more efficient for keeping track of the relationships between those notes than a tree directory filled with docs.

Re:Gemini 4 Argon on Tuesday October 13, 2026 @04:06PM (#50077864)
Irrelevant. The target audience was Google VPs.
Re:11 year old unsoundness bugs found in Rust borrow checker on Tuesday October 13, 2026 @08:49AM (#50070862)
> Rust cannot fail, it can only be failed.

“Please build the perfect system. Because I specified ‘perfect’ any flaws are due to your incomplete implementation, not my awful brief.”

If the person writing the rust spec was a client I would have to drop them because they would be setting themselves up for dissatisfaction.

Re:GPT‑6 and Intelligent UI for everyone on Wednesday October 07, 2026 @06:16PM (#49999519)
It's okay to be able to grasp what is important, and move on. You have a finite amount of time to spend.

For me, I try to spend as little time in QuickBooks as possible.

Re:GPT‑6 and Intelligent UI for everyone on Wednesday October 07, 2026 @05:46PM (#49999230)
> Feels like I'm being condescended to and treated like a child.

My most used prompt in the last week is probably "Explain this concisely and simply, like I am a child".

I spent years dumbing things down and creating visuals to support it for decision makers. I am frequently asking ChatGPT to do the same for me.

Re:Agents don't need memory. They need documentation on Sunday October 04, 2026 @12:23AM (#49950607)
The brain has mechanisms that can organize experiences without requiring that relationship to be expressed as a sentence.

Like, you can quickly lookup related ideas based on what came before and after, causes, effects, just like calling relationships a graph database.

Documents can’t be queried efficiently like that, you need a database.

Re:Opus 5.5 is good at explainer videos on Friday September 25, 2026 @09:09AM (#49844152)
What about teaching employees the basics of a job?

I love using notebookLM for orientation stuff.

Re:A pharmacy chain in Vermont implemented AI for efficiency on Wednesday July 29, 2026 @10:28PM (#49105485)
I think that’s the idea behind those centralized pharmacies like Amazon health. Easier than putting a pill dispensing assembly line in every CVS.
Surveillance and locks are both imperfect solutions to the trust problem.
I’m not sure an absence of surveillance is what creates “humane trust”. I’m certain we had locks on doors and security guards before the internet.
Re:CarPlay Is Additive on Friday July 03, 2026 @08:03AM (#48774014)
Aux cables were ubiquitous when cars were rolling speakers. I wouldn’t buy a car without an aux input, I didn’t care how many cds it could hold.

Now that cars are rolling screens, I wouldn’t buy a car without screen mirroring for the same reason.

Re:Sprinkling self-doubt on ChatGPT on Friday August 22, 2025 @09:06PM (#44991946)
We should probably be aiming for some stress, but not too much. See *Yerkes–Dodson*

https://en.wikipedia.org/wiki/Yerkes%E2%80%93Dodson_law

You’re right, I had that backwards
Can we put energy in to speed it up?

That way a day can be 24 hours exactly instead of 23 hours, 56 minutes, 4 seconds, etc...

I guess only Nixon could go to China?

When the US government adopts an employment policy private companies usually follow. Either because they do business with the government and have to comply, or it affects so many people there's a shift in norms.

Re:IMG_0416 on Sunday November 10, 2024 @08:05PM (#42103759)
It is! I've often wanted to search for something that happens in a video instead of just the title, description, and keywords.
It's probably safe to assume that if you're looking for a fork of Firecrawl, you already know what Firecrawl does.
> Karen-like drivers who will eventually start threatening other people to report them,

That's an argument for automating the system, taking the biased human actor out of the process.

We’re on the same page.

You’re acknowledging that the safety record is just a proxy for safety. That’s what I’m saying too.

I would argue that using the number of incidents as a proxy for safety ignores the trend.

So your assesment of how safe the plane is doesn't depend on anything except Boeing's process? For example, you don't care about the number of accidents per operating hours?
If you want to be pedantic, you're conflating 'having a safe record' with 'being safe.' It's just an indicator.
> nobody seems to care?

Clearly people care. They are literally checking every bolt for proper torque. When they found a few undertorque bolts it becomes a national news, it gets shared so much it ended up on Hacker News, and you cared enough to comment.

"end up being" is not enough of a disclaimer for you?
737 Max is going to end up being one of the safest planes to fly in.

Every inch of every plane combed over, every engineering decision re-examined, every system and sensor double-checked for redundancy and reliability, the pilots alert for any sign of problems.

Re:Mechanical Watch on Monday December 11, 2023 @10:15AM (#38601340)
For me it was an easy choice. First, I was starting to worry about heart health. Second, my mechanical watches could be sold for more than I paid for them.
Re:OpenAI board in discussions with Sam Altman to return as CEO on Saturday November 18, 2023 @09:11PM (#38327684)
Who sacks the person who sacks?

Older »