Hacker Newsnew | past | comments | ask | show | jobs | submit | bredren's commentslogin

Yes. Gemeni's web interface was atrocious even when the model was the best frontier.

And codex still uses phrases and syntax in prose ostensibly for the user as though they forgot people are actively reading this stuff.

Product is unquestionably where Anthropic excels. It is what carried it through periods where its thinking model lagged.


I had a similar thing happen where I was looking to recover funds from unexpected extra usage charges and got went through an identical experience.

I realize the company barely has time to cash checks, but failing to handle small fry reasonable charge disputes should be handled appropriately.


It is satisfying to see someone hacking on deprecated hardware and software also is keen to look forward into Vision Pro.

Claude Code and Codex are not SaaS products in the traditional sense.

Can anyone point at the critical vulnerabilities already patched as a result of mythos? (see 3:52 in the video)

For example, the 27 year old openbsd remote crash bug, or the Linux privilege escalation bugs?

I know we've had some long-standing high profile, LLM-found bugs discussed but seems unlikely there was speculation they were found by a previously unannounced frontier model.

[0] https://www.youtube.com/watch?v=INGOC6-LLv0


- The OpenBSD one is 'TCP packets with invalid SACK options could crash the kernel' https://cdn.openbsd.org/pub/OpenBSD/patches/7.8/common/025_s...

- One (patched) Linux kernel bug is 'UaF when sys_futex_requeue() is used with different flags' https://github.com/torvalds/linux/commit/e2f78c7ec1655fedd94...

These links are from the more-detailed 'Assessing Claude Mythos Preview’s cybersecurity capabilities' post released today https://red.anthropic.com/2026/mythos-preview/, which includes more detail on some of the public/fixed issues (like the OpenBSD one) as well as hashes for several unreleased reports and PoCs.


That OpenBSD one is exactly the kind of bug that easily slips past a human. Especially as the code worked perfectly under regular circumstances.

Looks like they've been approaching folks with their findings for at least a few weeks before this article.


While not entirely unrelated, Linux also had a remote SACK issue ~ 6 years back.

So if this Mythos is just an expensive combination of better RL and the original source material, that should hopefully point out where we might see an uptick in work ( as opposed to a novel class of attack vectors).


It is a big claim without the source and prompting.

I had thought this, but my experience initially was that performance degradation began getting noticeable not long after crossing the old 250k barrier.

So, it has been convenient to not have hard stops / allow for extra but I still try to /clear at an actual 25% of the 1M anyhow.

This is in contrast to my use of the 1M opus model this past fall over the API, which seemed to perform more steadily.


I also didn’t like the example. How may people know the difference between chocolate quality?

Some for sure, but I wouldn’t count on it for interpreting price elasticity.

I still believe in the premise because of the action on Facebook buy nothing groups.


Did you miss the part where people would pay more for the "high quality" chocolate when the price wasn't zero?

Guess so _/ \_

This is not an unknown issue at the fruit co.

Can anyone speculate on any rational if not good reasons for not solving this problem yet?


I don’t work at the fruit co but since you asked for speculations. Mine: the fruit co designers are still designing a nice interface to show the overflow, because they obviously think that the Windows tray overflow looked inelegant and are still searching for the ideal UI. But the designers themselves don’t have a lot of menu bar apps so they don’t think it’s a priority.

Or perhaps the teams at fruit co found a way to claim that their overflow is an innovative new feature and not copied from some other designs.

While they do a ton of good work, they do love to claim everything was first invented by them.


Probably the same response I just saw someone reply with in this very thread:

"You shouldn't have so many utilities running"

It's the go-to Apple user response to anything the OS doesn't support or does poorly: "Why would you want to do that?"


Windows has always baffled me with the system tray icons it is too cluttered. I grew up with a tricked out Linux desktop so I understand the need to customize. But most of the time you do not need that.

I believe a VPN should stay hidden if it works, no need to have it visible.


> I believe a VPN should stay hidden if it works, no need to have it visible.

Which is fine if you only have one VPN client or one VPN network and you don't need to turn it on/off or change it regularly.

My current day job has one VPN client but five different networks.

At a previous job I had two different clients I would need to switch on and off.

It is very on-brand with Apple though that there is one right way to do things, and everyone else either needs to change the way they do things or go elsewhere.


I disagree with this one. If a VPN is important, I want to see that it is still connected and hasn't crashed.

That’s the company response but I’m definitely not the only long-term Apple user whose go-to response is a sympathetic nod followed by a long rant about Tim Cook and his contempt for software engineering.

Considering that I need a good dozen utility apps to override absolutely bonkers macOS design descicions there is no way around that.

TBF, there isn't a computer on earth that will solve that problem perfectly. At some point, "you shouldn't have so many utilities running" is perfectly acceptable advice.

No, because their icons can simply be collapsed into a disclosure control.

"You'll run out of memory eventually" was my point.

That's the standard apologist response to ANY defect you point out in anything, or any question they don't know the answer to but still want to bloviate about.

See: Stack Overflow


The upcoming MacBook Pro (late this year) is rumored to have a hole-punch camera: https://www.macrumors.com/2026/02/24/touchscreen-macbook-pro...

It‘s reasonable to assume that menu bar items will be rendered differently as well, to accommodate for Dynamic Island (which changes its width as needed).


Well I mean, recently because they have no idea how to make good UIs, and have not read their own enormously detailed (and excellent) Human Interface Guidelines tomes from 10, 20, and probably 30 years go, and have basically regressed to barbarism.

But before that relatively recent fall-off-a-cliff event (whatever it was that caused it, most of us will never know), it was pretty clear that they didn't want to implicitly endorse the lazy/anti-user/Windows-equivalent-UX antipattern of having apps that intentionally made themselves accessible only from a menu bar icon.

I hate the App Store shite that goes wildly too far the other way, but I don't quite understand wwhy they couldn't figure out a way to enable the menu bar widget API in a way that failed if your app didn't also have a way to open via all the normal ways (double-clicking the icon in /Applications, asking Siri to launch it, etc)


> they didn't want to implicitly endorse the lazy/anti-user/Windows-equivalent-UX antipattern of having apps that intentionally made themselves accessible only from a menu bar icon.

The single biggest complaint I had when I switched it to Mac was lack of this feature. Still miss it. .


> and have not read their own enormously detailed (and excellent) Human Interface Guidelines tomes

This seems to also apply to all new UIs produced by apple in the last 5 years.


They think you're holding it wrong.

Pelican is drafting rear peloton

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: