What you actually cannot afford is have email clients that are easily hacked via email. Like Outlook.

Everyone just wants more VRAM... GPU performance itself is less important in an environment where the trendline is larger sparser models.

I don't.

The creator of the pi harness did say that the claude system prompt changes often enough that he had to build his own tooling to try and improve reliability and repeatability.

I'm not convinced the differences in roles I've observed are claude personality differences as much as humans giving different prompts.

I would be fired and sued into oblivion if I did this.

Previously you claimed "I wrote a line of code by hand saturday (client blocks AI) and it was weird. I hadn't done it in months." now just 90%. What happened? Did you decide to turn the hyperbole down slightly because nobody was buying your bullshit?

This from the same handle that previously said "A rough estimate is that the best LLMs do software engineering about as well as IQ 150 software engineers do it. That makes them better at it than the vast majority of people who write code."

I have AI spit out code occasionally and rarely accept the output as-is. Usually you can replace it with something much better and simpler. It'll often compile and run but I can't even imagine the technical debt you vibe coders are racking up.

This discussion on their inclusion in Mozilla is quite revealing:

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fbugzilla.mozilla.org%2Fs...

Elsewhere it's claimed that they used a FIPS HSM to generate their keys.

I did not see who their auditor was.

Did eMudhra lie?
Did the auditor lie?
Are FIPS HSM's backdoored?
Do we have another Dual_EC_DRBG situation on our hands here?

If nobody is lying we could have just seen a huge cyberwar or false flag averted. Essential crypto gear could be totally buggered still.

PS Remember when Mozilla gave CACert absolute hell for years when they were tremendously more competent?

We just migrated to a hosted server for our main point of sale/inventory system. Not because we wanted to, but because there was literally no other choice. The in-house server was end of life, and the vendor just refused to sell us a new one. There's nothing - we have looked - that can replace it, open source or otherwise, without extensive customization that would take several years and require a full time staff of programmers, plus the active cooperation of the national franchise that would not be forthcoming, at a cost of millions.

And it's not an upgrade. We've had more offline time in the last two weeks that we did in the last year with the in-house server, and there are some very important functions that are, at best, unstable and almost unusably slower. They can't even get the login screen to work right.

And it costs us about 50% more than the platinum support contract we had before.

That's not an upgrade, it's a clusterfuck.

If there were an open source point of sale/inventory package available that was capable of replacing our current proprietary system, we'd be on it in a second. Until there is, and there never will be, we're stuck.

I've been using two frontier models, each acting as the other's adversarial reviewer, and it's worked out quite well. I've been able to write production code since the beginning of the year, and I've noticed a marked improvement in that production code. Over the last three months, the models really are getting better at the frontier.

The one caveat is that AI will try to solve every possible edge case with an overly robust solution that adds a lot of computational, codebase, and unit-testing overhead when it is unnecessary. It doesn't know how to manage the probabilities of edge cases popping up and what to do if those edge cases do happen. There are some edge cases in my business where they are vanishingly unlikely to happen, and the correct response is to log the problem, deny access, and fire off a quick message to operations. The AI models will give you an overbuilt monstrosity. It really wants to hit a gnat with a sledgehammer.

I've found that the biggest problem I have to solve with AI is teaching it to act like my favorite designer and engineer, Antoine de Saint-Exupéry, who used to say, "Perfection lies not in asking, 'Is there nothing left to add?' but rather, 'Is there nothing left to take away?'"

How do you decide to trust code you write?

How do you decide to trust code someone else wrote?

Tests. Design discussions. Code reviews. Code review questions. Design Questions. "What about this problem?" Real validation testing.

I was in a funny situation last week. Two different humans told their claude agents to fix the same bug. Claude came up with two _slightly_ different implementations.

I told _my_ claude agent to code review both implementations - using a new prompt I'm working on to try and automate away a lot of the code review value I was adding manually.

Same model; same bug, two implementations, and the code-reviewer agent found different bugs in both PRs.

We ended up blending the best things from both PRs into the one that shipped.

I am really confident that the code I ended up shipping here is better than what our specific humans would have done without using AI, and 10x-100x faster.

I am spending more of my time telling agents how to think like me, what to look for, what tradeoffs to make, etc. They wont do those things unless I tell them to. They don't always do those things perfectly.

(Neither do the real human software engineers who work for me.)

Use a paid subscription to a frontier model. Make sure it knows how to run your build and run your tests.

If you're worried, have it write tests. Look at the tests; see if they look like the right tests.

Tell it to ask you questions if it needs to resolve something ambiguous. Have it explain what it thinks it should do to the code. Ask it probing questions if you aren't sure it understands. Ask it about design tradeoffs.

I have yet to find a problem where AI generated code doesn't get me at least 90% of the way there. Multi-domain, multi-language, multi-platform. That includes stuff I was doing a year ago. It's gotten dramatically better in 2026.

If someone has an example of a software engineering problem that AI actually did really badly on, I would like to know the model and see the prompts. I'm not saying it doesn't exist. I'm saying I haven't seen it yet, and I'm really curious.

I've done software development professionally since the 1990s. I started coding on the Apple ][ in 5th grade. I've had AI do the wrong thing, or misunderstand something, or miss something that seemed obvious to me. But on all of those same criteria; I've had human software engineers do far worse, far more often.

The ones overloading my web server are, without any doubt, AI bots. Most of them honor the robots.txt file. Amazon's, however, do not, and constitute more traffic than everything else combined (and it get a fair amount of legitimate traffic per day).

It would have to reproduce the light field to sell the illusion. Otherwise would be obvious especially when walking around.

Hello fellow Frame owner.

I bought a Frame 65 just because of how beautiful the display was in Art Mode, and I want pretty art in my house for the 90% of the time the TV would otherwise be off.

I use the "Art Mode", but only with my own sideloaded content.

One thing that stinks is that sideloaded art doesn't support the full Art mode experience; specifically no Metadata. I seeded the display with lots of public domain images of classical and romantic european art (my taste - and why I never considered a paid art mode subscription), and I'm not cultured enough yet to just remember what all of it is. So I wish I could get the tv to show metadata sidecars for my custom images.

The Samsung firmware is garbage. I hate it whenever I accidentally hit the remote button for voice assist or AI assist or any of that garbage.

But the way the display looks when showing paintings is too good to pass up. If i could get a firmware that took out home screen nonsense and disabled 80% of what the TV did, but added metadata display for sideloaded images... it would be perfect.

Why is Palmer Luckey a "creep"?

Are you a TikTok girl or something?

Have you actually looked into his relationship habits?

He is a huge anime and retro video game nut. He married a cosplayer. (He did a little cosplay himself). They've been married since 2019 and have a kid together.

He won at life, basically.