I agree; unfortunately when I brought up that they're losing before I get jumped on demanding me to "prove it" and I guess pointing at their balance sheets isn't good enough.
I won't argue with your point; both Anthropic and OpenAI name their models poorly, and it is hard to follow unless you're already following it.
"Sonnet" only makes sense relative to other things but not by itself. If you don't know those other things, it is difficult to understand.
But, if you were asking (and I'm not sure that you are): "Sonnet 4.6 is a cheaper, but worse, version of Opus 4.6 which itself is like GPT-5.3 Codex with Thinking High. Making Sonnet 4.6 like a ChatGPT 5.3 Thinking Standard model."
Unless you're interested in this type of stuff, I'm not sure you really need to. Claude, Google, and ChatGPT have been fairly aggressive at pushing you towards whatever their latest shiny is and retiring the old one.
Only time it matters if you're using some type of agnostic "router" service.
Student Loan interest rates in the US can be as high as 9-13%. The government can borrow at 3.36% which even if we assume a 20% overhead is 4.03% to the borrower. Other countries/governments do a scheme similar to this, and it makes repayment realistic.
I'm certain someone will respond telling me the difference between Subsidized, Unsubsidized, PLUS, and private loans which completely missing the point: There shouldn't be a private entity that needs to turn a profit on the backs of students begin with, it is immoral. If you remove the private for-profit entity, the loan-type distinction goes away.
It isn't uncommon to read stories from people, who graduated and are in good jobs, and had no gaps in repayments that are now on 300%+ of their original borrowed amount.
Your argument kinda falls apart with the observation that private loans often have waaaay lower interest rates than federal ones... Typical private loans start in the 3% range, vs 6.4 for the lowest Federal ones.
Can you think to the "3% range" private student loans? But regardless, my "argument" is for lower, fairer, interest rates that students are actually expected (and can realistically) repay. One where profit isn't made.
Who makes the profit is largely irrelevant; the students suffer the same either way.
Current Opus 4.6 would be a huge achievement that would keep me satisfied for a very long time. However, I'm not quite as optimistic from what I've seen. The Quants that can run on a 24 GB Macbook are pretty "dumb." They're like anti-Thinking models; making very obvious mistakes and confusing themselves.
One big factor for local LLMs is that large context windows will seemingly always require large memory footprints. Without a large context window, you'll never get that Opus 4.6-like feel.
Yep; and all Apple fans ever say is "report feedback!!!" but what is the point when seemingly Apple never gets to their backlog of bugs/broken features? I mean, sure, some big stuff gets fixed, but there is a lot of stuff broken going on years they haven't even touched.
Feedback is more or less a black hole, for the most part. It's rarely paid attention to by a human, and is treated like telemetry. If you want something fixed, it needs to get into the press or go viral.
I’m sure there are countless examples to the contrary, but I recently submitted feedback regarding an issue that I was experiencing in Final Cut Pro. Within a week, a member of the Final Cut Pro team contacted me and asked for a copy of my video editing files so they could replicate the issue. I sent them the files, they confirmed the issue, and the issue was fixed in the next release.
> Yep; and all Apple fans ever say is "report feedback!!!"
I'm trying not fall into "No True Scotsman" but... It should be common knowledge at this point that Apple Feedback is a blackhole of despair. "Please attach a sample project" seems to be the go-to, even for things were that makes no sense. Same with attaching debug/diagnostic logs. I understand the value of all of those things but even people who have jumped through all the hoops get ghosted and/or their issue is never addressed.
Currently I would not waste my time on Feedback and it's sad because even if Apple reverses course it will take a lot to get the people who they should most want creating Feedbacks to create them.
Those are probably anonymous employee accounts not fans. I don't know if anyone would be enough of a power user fan to tell someone to file a bug report.
OpenClaw sits on top of a physical machine/VM you control, you give it (hopefully) limited/sandbox access to that machine's resources to act like-you, and it does useful things. OpenClaw's user interface is just a gateway, and is only as useful as whatever the machine/VM has under the hood.
So the "setting it up yourself [on a VM/machine you control]" is kind of core to the whole idea being useful, you take that away, and it is just another Chat-Bot? Making it more of an ChatGPT/et al competitor rather than OpenClaw.
Installing it on your PC or laptop puts your personal data and ISP subscription at risk, while installing it in a hosted VM yourself requires a bunch of Linux security and networking knowledge or else you'll get pwned pretty much immediately (https://youtu.be/40SnEd1RWUU). So this service is giving you a VM already set up with a security baseline.
Is that what this does? All the link takes us to is an empty website about "Kimi Claw."
The entire crutch of the "Claw" concept is being able to directly reconfigure the VM/Machine to be "your" environment (to a point). A blank VM with nothing configured on it, is as useful as a cardboard bathtub.
Ultimately this link is a terrible intro to whatever this is.
Hm, that YouTube video made me think a bit, sure if you put it all like that, it does feel like a lot of stuff to get right, but whenever I do it, it takes about 30 minutes to lock down the firewall, do some port-scans to verify, punch a VPN through and hide SSH behind it. That way you're already protected from 99.9% of attacks, and then hope that that last tenth of a percent won't stumble upon you, and also that the VPN is secure enough, though I guess if that is breached it's not only you who's fucked. Also you need to look out that Docker doesn't destroy your firewall. I don't know, it doesn't feel like that much work, right? Maybe I'm just blind to it.
What you and I consider routine work, someone who works with mostly Webdev or might consider extremely difficult. There a lot of programmers who have never used a Linux shell, or know much about networking beyond TCP, or used Linux before outside of a uni class 20 years ago.
3rd Party Keyboards exist, but they don't have the same rights/abilities as Apple's native keyboard, directly resulting in some features/functions being impossible to implement.
- Apple's keyboard has richer context; third-party keyboards are largely limited to the text proxy (surrounding text/selection) and cannot inspect the host app/see the underlying context of where the text is used/what for.
- No caret/cursor control; cannot reliably set the cursor to arbitrary positions
- No custom dictation pipeline inside the keyboard; must rely on system dictation (or move voice input outside the keyboard).
- No camera access from the keyboard extension; blocks scan-to-type (OCR), QR/barcode capture, etc.
- Can do basic text expansions, but cannot implement privileged automation/macros that interact with other apps/UI or run broadly in the background.
- Due to aggressive lifecycle + tight resource limits constrain large language models/dictionaries; cloud sync/personalization
That definitely explains SOME of why SwiftKey is worse on iOS, but it doesn't explain much of it. It just seems like Microsoft never got it to feature parity.
reply