dmazin

Putting the makers of an OSS harness in the same moral group of Palantir and Anduril is laughably silly.

> GPT-6 Luna being half the price of GPT-5.6 Luna is a really big deal.

Is it? It was already too cheap to meter for me. Luna 6 is actually worse on some benchmarks than 5.6. I’d have loved improved performance for 2x the price than ~equal performance for 0.5x the price.

Per the benchmarks in the post, Luna 6 is at best a couple points superior to Luna 5.6 and (unless I’m reading it wrong) xhigh has actually degraded in quality?

I was hoping for a serious Luna upgrade. It was already cheap enough. This feels more like a price reduction than an upgrade.

That said, if the new Luna is able to handle ultra mode and subagents v2 in codex cli, then at least that’s a win.

It seems like they are. I mean, this is similar in performance to Fable (ish). It seems like more focus on making existing capabilities more accessible.

I think this refers to the trial.

This feels like a solved problem entirely: Kagi. I don’t know their secret sauce but the results are very high quality.

It would appear you are missing the High Barnet branch of the Northern line.

Have you considered that he has had to walk back the cuts because of the outcry?

The point of a graduation speech like this is to get students hyped up about themselves and their future. Surely you see the merit in, amongst a backdrop of a horrible job market, telling students that they have, inherent in them, the stuff of greatness, just as people did (checks watch) 3 years ago before vibe coding?

As a few commenters already pointed out, IME enterprises aren't paying for subscriptions. They're paying per token.

But also... is this shit AI written? I'm so tired of this.

Yeah, that's definitely my favorite book about Apple/Steve Jobs.

If you want more on this, I recommend Steve Jobs and the Next Big Thing by Stross. I’m not sure, but it might be the only extensive book about Next other than this new one.

Though it’s essentially a long hit piece. The author really had it out for Jobs.

In fact it’s a completely uncharitable book now that I think about it. Hopefully this new book will be a lot less biased.

Yeah, 1/1 is one song I always keep downloaded to my phone for this reason!

Thanks for that article, love to read about well intended design being poorly received.

Was it in that kind of touristy area filled with children? I didn’t think to go in.

Is this a regular thing?!

Has anyone actually heard Eno at the airport? What is it like? Does it actually calm you?

I got sick of the inconsistency caused by Anthropic tinkering with Claude Code and had canceled my 20x. My plan was to switch to Codex so I could use it in Pi.

I am specifically talking about switching because of the harness, not model quality. Anyone else match my experience?

I wonder how many other people recently did the same. It would be prudent of Anthropic to let people use Pro/Max OAuth tokens with other harnesses I think. Even though I get why they want to own the eyeballs.

Agreed.

That said if this bothers you I highly recommend not looking up how many Space Shuttle missions are classified.

Constraints can lead to innovation. Just two things that I think will get dramatically better now that companies have incentive to focus on them:

* harness design

* small models (both local and not)

I think there is tremendous low hanging fruit in both areas still.

why do half the comments here read like ai trying to boost some sort of scam?

Maciej now has a Mars newsletter, which I obviously subscribed to immediately: https://mceglowski.substack.com/

I didn’t even have a strong interest in space before the dude started writing about it. Maciej could write about literal rocks and make it worthwhile to read.

Well, it should only update what it says: security updates (from official Ubuntu sources) unless you change the configuration.

This is a lot less of a story than it seems.

It makes it sound like a rogue AI hacked Meta.

Instead, the "wild" thing here is that someone let an agent speak on their behalf with no review. The agent posted inaccurate instructions which someone else followed.

Those instructions lead to a brief gap in internal ACL controls, sounds like. I'm sorry, but given that the US government gave 14 year olds off incel Discords full access to Social Security data, this is not shocking by comparison.

To be clear, it is dumb and rude to let an agent speak on your behalf _without even reviewing it_.

This will eventually lead to a bigger snafu, of course. Security teams should control or at least review the agent permissions of every installation. Everyone is adopting this stuff, and a whole lot of people are going to set it up lazily/wrong (yolo mode at work).

Me: “I want to wash my car. The car wash is 50 meters away. Should I walk or drive?” Opus 4.6, without searching the web: “Drive. You’re going to a car wash. ”

If you can get your hands on it, I recommend Other Networks: A Radical Technology Sourcebook by the same author. She covers barbed wire as well as many other ways to communicate. The book itself is gorgeous.

While NYT etc mostly stand back as the U.S. crosses the rubicon, what started as a freaking gadget review site seems to employ nearly all the journalists with actual gusto left in America.

Wow – thank you for that.

Ah, rest in peace. I recommend you watch Computer Chronicles. It’s a treasure. If nothing else, it’s a solid trip down memory lane.

Yeah, the audio engineer then talked my ear off about networking!

I highly recommend anyone to look up how PTP works and how it compares to NTP. Clock sync is very interesting. When I joined an HFT company, first thing I did was understand this stuff. We care about it a lot[1].

If you want a specific question to answer, answer this: why does PTP need hardware timestamping to achieve high precision (where the network card itself assigns timestamps to packets, rather than having the kernel do it as part of TCP/IP processing)? If we use software timestamps, why can we do microsecond precision at best? If you understand this, it goes a very long way to understanding the core ideas behind precise clock sync.

Once you have a solid understanding of PTP, look into White Rabbit. They’re able to sync two clocks with sub-ns precision. In case that isn’t obvious, that is absolutely insane.

[1] So do a lot of people. For example audio engineers. Once, an audio engineer absolutely talked my ear off about ptp. I had no idea that audio people understood clock sync so well but they do!

Ask ChatGPT to generate product or business names for you. A lot of them contain the word “loom”. I know someone who recently vibe coded a thing and had an AI name it and it also has “loom” in it.

I’m not suggesting that that’s what happened with this project. I have no idea how much AI they used for any of it. But “loom” seems to appear in a lot of ChatGPT generated product names lately.