Sorry, english is not my first language. Itss disabled on our work API instances so my use is strictly personal because it was available on Cursor (my account is pre-acquisition). I have extensive AI experience across nearly all models, open weight and frontier and closed, I am very aware of how they talk or output. This is high/xhigh, via Cursor, not via X Grok/UltraGrok if that make different.
Grok 4.6/4.7 feel like you are less tokens because Cursor gives you double the quota subsidising Grok specifically - there is a separate Grok/Cursor model quota bar in addition to non . Its actual token drain if you are paying API rates is generally far higher than Fable, Astra et al. You are not using less tokens.
It is weird in some reward way, it will frequently, at least for me, complete the task in the laziest way possible,Technically it is done, but that's about it.
For example a user registration system ended up with users being able to log in as anyone because the authn was a cookie set with the user's pkey id, unsigned. Nearly anything it outputs technically works but if you give it a GLM or DS pass you will find dozens to hundreds of vulnerabilities of varying hilarity (Fable refused, Opus refused).
I have seen it write SQLI-vulnerable code, asked it to review a file without saying what's wrong and it did not catch it (fresh session, single file @-tagging in harness). Again, technically, the code works and does what the prompt asked, its just handing in some of the laziest copied homework I have ever seen.
If you specifically call out to use prepared statements it will typically put a plaster over this however I have also seen Grok 4.6 use prepared statements by concenating the user input raw into a statement then executing it with no ? or named replacements, rendering it somehow a SQLI-vulnerable prepared statement. One med it added a OR 1=?, replaced the ? with another 1, then said OK you are using prepared statements now.
We received repeated, constant heads up to move our data by the first AZ much less second. The problem is that nobody is storing data in Bahrain unless there are data residency requirements for it.
nobody wakes up one morning and chooses to launch instances, CDN or S3 and would choose Bahrain as that without a requirement to, we were contractually and legally forbidden (in the middle as a vendor) to copy even encrypted data where we don't have the key out for redundancy, so the best we could do was tell our subcustomers to download all of their buckets to their office or some employee laptops at their office
And it's not a lead spam database. It's used for connecting with subject matter experts if you want to ask questions or ask for feedback/improvement on an application, or for research on anything
Thank you for pointing that out, and I apologize for the lack of clarity in my earlier statement. In this process, we utilized ChatGPT to process and generate the related keywords.
I think the confusion comes from the fact that the linked page is not a text article, its actually an AUTOPLAYING VIDEO for a small news station. There is not actual text content
IIRC, if you have a private key you can be able to force a revocation regardless of what the owner wants. In some such as Let's Encrypt it is fully automated.
If this is a repo private, you should be realize it with a private CA that you import or is on every corp machine.
Baseline Requirements force a revocation within x hours on key disclosure.
I didn't know about CA/Browser forum and the Baseline Requirements. Thanks, will check it out!
// Edit: Relevant section:
The Subscriber Agreement or Terms of Use MUST contain provisions imposing on the Applicant [..] the following obligations and warranties:
[...]
Protection of Private Key: An obligation and warranty by the Applicant to take all reasonable measures to assure control of, keep confidential, and properly protect at all times the Private Key [...]
> Looking at digicert[1], if a revocation request is submitted, the owner must approve it. What happens if I just don't approve it?
So in this case, this is the happy-case where you as the owner wish to simply realize the cancellation a cert that you are no longer using.
A different workflow applies, such that you have the private key you instead send a POST to 'https://problemreport.digicert.com/api/keys/compromised' with the private key in the JSON body and it will be queued. It is mandatory Baseline Requirements wise to cancel the certificate within 24 hours in the compromised case - usually instant if the pk matches cert - with the expectation that of course the owner will not go this route.
We haven't had any issue getting all of our staff -- nontechnical users alike -- on yubikeys. As part of education we skip the PKI portion and just point out that it is "like your physical house key. You plug it in and touch it to turn the key to unlock"
For business use, handing out yubikeys is completely reasonable - if they're lost or broken, the company is the authority and they can (and do) just re-issue a new one and work goes on with only a temporary interruption. They're easy to explain, fast to use, provide practical security, and are simple to recover in case of total failure.
For personal use, that equation is wildly different. Google isn't going to let you attach a brand new key, you've just lost your account forever because it rained.
Yes, and our staff uses ssh keys (generally but not always without issue) and a commercial MFA app. It's one thing to get this stuff used in a controlled environemnt where you have a help desk or administrators who can do a lot of the setup. You just hand the employee their YubiKey or smart card and say "use this."
Trying to imagine your grandmother setting it up herself to be able to log in to her Facebook is another matter, and why these things have never worked for the general public.
You probably use certificates and a company PKI to manage them. No need to stress if one is lost or locked, just revoke and whip up a new certificate.
At home Yubikey is probably synonymous to FIDO not PIV/PKI. No whipping up a new one if you lose it. You better have 3 of them enrolled at any time, and have at least one stored off site.
We enroll them as standard fido/webauthn - I hate the other modes.
I agree it requires significantly more work when you can't just call the locksmith for a new one -- IT -- if you lose one on your personal account you can only go get the spare key hidden under the doormat, a printed code in your safe, or lose the account.
Grok 4.6/4.7 feel like you are less tokens because Cursor gives you double the quota subsidising Grok specifically - there is a separate Grok/Cursor model quota bar in addition to non . Its actual token drain if you are paying API rates is generally far higher than Fable, Astra et al. You are not using less tokens.
It is weird in some reward way, it will frequently, at least for me, complete the task in the laziest way possible,Technically it is done, but that's about it.
For example a user registration system ended up with users being able to log in as anyone because the authn was a cookie set with the user's pkey id, unsigned. Nearly anything it outputs technically works but if you give it a GLM or DS pass you will find dozens to hundreds of vulnerabilities of varying hilarity (Fable refused, Opus refused).
I have seen it write SQLI-vulnerable code, asked it to review a file without saying what's wrong and it did not catch it (fresh session, single file @-tagging in harness). Again, technically, the code works and does what the prompt asked, its just handing in some of the laziest copied homework I have ever seen.
If you specifically call out to use prepared statements it will typically put a plaster over this however I have also seen Grok 4.6 use prepared statements by concenating the user input raw into a statement then executing it with no ? or named replacements, rendering it somehow a SQLI-vulnerable prepared statement. One med it added a OR 1=?, replaced the ? with another 1, then said OK you are using prepared statements now.