Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
_Val_
Admin
Admin

ChatGPT on CheckMates?

We are currently looking into the possibility of "inviting" ChatGPT into the CheckMates community. While this idea is still in discussion, and there will be some time before we can make it happen, I played with the technology a bit.

Here are some results, and they are not bad at all:

VAL: How do I check status of Check Point gateway cluster?

ChatGPT: To check the status of a Check Point gateway cluster, you can use the "cphaprob stat" command in the command line interface (CLI) of the management server. This command will display the current state of each cluster member, as well as information about the cluster's virtual IP address and sync status.
You can also use the SmartConsole GUI to check the status of the cluster by navigating to the cluster object and viewing the "General Properties" or "Cluster Members" tabs.

Additionally, you can check the status of the cluster by using the Check Point API with the command "cphaprob list" and get the cluster member state.


VAL: I have installed "Any Any accept" policy on my Check Point security gateway. but I cannot ping it. How should I troubleshoot the issue?

ChatGPT: If you have installed an "Any Any accept" policy on your Check Point security gateway but are unable to ping it, there are several steps you can take to troubleshoot the issue:

Verify that the security gateway is properly configured and reachable. You can do this by checking the gateway's IP address and connectivity, and by testing connectivity to the gateway from the management server or another host on the network.

Verify that the "Any Any accept" policy is properly configured and installed on the security gateway. Make sure that the rule is in the correct position in the rulebase and that it is not being overridden by a more restrictive rule.

Check the gateway's firewall log to see if there are any dropped packets or denied connections that may be related to the issue.

Check the gateway's routing table to ensure that the gateway is properly configured to forward traffic to the correct destination.

Ensure that the security gateway is configured to allow ICMP traffic through. By default, the "Any Any accept" policy does not allow ICMP traffic, you need to add explicit rule for ICMP traffic.

Check if there are any other security measures like Network Address Translation (NAT) or VPN that could be blocking the ICMP traffic.

If all the above steps have been tried, you can also try to reboot the gateway and see if that resolves the issue.

By following these steps, you should be able to identify and resolve the issue that is preventing you from pinging the security gateway.

 

And here is my favorite:

VAL: create a sonnet about Check Point firewalls and stateful inspection

ChatGPT:
Oh Check Point firewalls, with stateful inspection,
Thou art the gatekeepers of the network domain,
Thy packets doth inspect with precision,
And only those that pass thy validation
Shall enter into the sanctity of our system.

Thy stateful nature doth remember
All communication, both in and out,
And with thy rules doth thee govern,
To keep the network free from doubt.

Thy protection doth not falter,
For thy threat prevention doth never sleep,
But vigilant doth thee remain,
Against the malicious who would creep.

Oh Check Point firewalls, with stateful inspection,
Thou art the guardian of our cyber perfection.

What do you think, folk? It seems, this tech could help us get answers faster, and also have some fun in the process. Let us know in the comments.

(1)
45 Replies
G_W_Albrecht
Legend
Legend

I see a problem in the hallucinated points showing up. If you get a mixed up step order or wrong parameters it could be a disaster. Also, all knowledge is from and from before 2021 only.

For help i have to be sure that i receive reliable information, so i will ask e.g. TAC that i trust a lot, CP documentation and experiences of fellow CheckMates. ChatGPT would be a funny oddball here, but not more - as it can not be trusted...

CCSE CCTE CCSM SMB Specialist
_Val_
Admin
Admin

I think you get it backward actually. All knowledge is after 2021 if I understand it correctly.

0 Kudos
G_W_Albrecht
Legend
Legend

No - only history up to end of 2021, that is the main reason for the ChatGPT wikipedia errors, blind for all after that date. If all knowledge is from after 2021 it could not do a sonnet...

CCSE CCTE CCSM SMB Specialist
0 Kudos
_Val_
Admin
Admin

One more thing from AI, a limerick this time:

There once was a network unsecured
Where hackers felt their skills were matured
But with strong passwords and firewalls high
They could no longer breach and spy
And the company's assets were secured

Bob_Zimmerman
Authority
Authority

The problem is ChatGPT gets technical stuff hilariously wrong. And it sounds 100% confident while doing it. For a fun example, ask it what LAST_ACK is. More often than not, it gives a response claiming the wrong end initiated a connection close, and the wrong end is waiting for the other to acknowledge.

TCP-state-machine-37.png

A bunch of programming forums have been inundated with people talking about how amazing it is and how the people posting are using it to write their code for them. Unfortunately, it also routinely gets extremely basic stuff like arithmetic catastrophically wrong:

IMG_7689.PNG

What is 3 + 5 again? Pretty sure it isn't 6.

G_W_Albrecht
Legend
Legend

I have read a long german article about explaining Wolf, Sheep and Cabbage to ChatGPT - it will always give a wrong answer no matter how often the solution is explained...

CCSE CCTE CCSM SMB Specialist
G_W_Albrecht
Legend
Legend

German scientists summed up the drawbacks well:

1. references/sources are often missing or wrong

2. you can not trust it (as it will make things up for you and present it as fact)

3. it is impossible to check if the output is fact or fiction if you do not know it already

CCSE CCTE CCSM SMB Specialist
the_rock
Legend
Legend

Well, everyone knows that German people are one of the smartest folks out there, so I think I will trust your scientists. I looked at some examples of this ChatGPT online and personally, just my opinion, I find it ridiculous : -).

G_W_Albrecht
Legend
Legend

I mentioned that they are german as it does not make much sense to give a link...

CCSE CCTE CCSM SMB Specialist
the_rock
Legend
Legend

I spoke some German back in high school, but not any more : - (. Anyway, regarding ChatGPT, personally, I would not waste second of my time on it.

0 Kudos
Bogdan_11
Explorer

Hehe, I remember when I was kid our teacher explained it to us and few of us were able to solve it. I heard this chat and other AI, are like similar to a 4-5 years old child. wondering how long will it take until these apps will replace humans  jobs entirely...

0 Kudos
Hugo_vd_Kooij
Advisor

For now I would put it in the funny section.

Too many examples of ..... questionable results.

 

<< We make miracles happen while you wait. The impossible jobs take just a wee bit longer. >>
Kurtis_Johnson
Employee
Employee

All the Technical Poets out there will be jobless.

(1)
_Val_
Admin
Admin

You made my day. Not with the way it rhimes now, they will not 🙂

Tony_Graham
Advisor

While I also am a fan of bleeding edge tech, as has been pointed out, there

is a little bit of peril here (cue Monty Python). However, occasionally we all run

into peril from human sources of information because every little nuance of

any situation may not be considered. So is s chatGPT inherently any worse? Maybe not.

I would recommend that the use of chatGPT be placed in its own section where someone

can 'take it under advisement' that they are in fact working with an AI and clearly mark any

information that comes from AI .I think flagging AI content is important and should be industry standard to prevent

AI from being used for disinformation but that's another topic.

Who knows, maybe chatGPT will become Checkmate of the year on its own merit?

Interesting to watch in any case.

the_rock
Legend
Legend

If ChatGPT ever became Checkmate of the year, I will close my account the same day...mark my words : - )

_Val_
Admin
Admin

We definitely need some more reactions in the community, not just kudos. I just cannot stop laughing. 

All you folks are anti-cyber bigots! Let's hope AI does not get any lawyering skills any time soon, because I see a huge, no, HUGE discrimination lawsuit right here, lol. 

0 Kudos
Bob_Zimmerman
Authority
Authority

Funny you should mention AI with lawyering skills. Somebody by the name of Joshua Browder started a company called DoNotPay to algorithmically fight speeding tickets and the like. A little over a week ago, he posted on Twitter that in February, they would have an AI practice law directly by telling someone what to say in court. This brought attention from a lot of lawyers. Bunches of them pointed out that if he did this, the person in the court and probably Joshua himself would wind up in jail.

To give an idea of the quality of advice DoNotPay gives, that particular case is for a speeding ticket. DoNotPay subpoenaed the cop who wrote the ticket. That is literally the worst thing they could have done pre-trial. In the US, if you take a ticket to court and cop doesn't show up (like if they have other work to be doing instead), the ticket is dismissed. The subpoena guarantees the cop will be there.

They also claim to practice immigration law, where even a small procedural mistake can result in the client being deported and banned for life. And they accept no liability if that happens. Super fun.

0 Kudos
the_rock
Legend
Legend

Never heard of that platform (DoNotPay), but reading about it, sounds, lets just say "interesting" haha. I mean, I cant imagine any normal judge that had practised law for a long time that would fall for that sort of stuff. 

0 Kudos
_Val_
Admin
Admin

You will be surprised. My Israeli buddy practising law for living claims ChatGPT, if properly trained on court materials, could be a HUGE thing.

0 Kudos
the_rock
Legend
Legend

I dont see that happening here in Canada, any time soon, if ever : - )

0 Kudos
_Val_
Admin
Admin

Uhhh, I see some worrisome messages here. It seems some of our experts suddenly feel threatened, lol.

It is a gimmick, folks, and we will treat it as such. 

0 Kudos
Bob_Zimmerman
Authority
Authority

Nah. I feel threatened by a Tesla driving near me. 😜

I find ChatGPT a little concerning because people are trusting it way past its point of reliability. As long as it's clear the AI may not know what it's talking about and following directions it provides verbatim might wipe your system, land your dog in the hospital, and/or bring about the apocalypse, it can be a decent tool for pointing out other avenues of research.

 

Separately, things like ChatGPT are a big part of why search engines are so useless these days. Algorithmically-generated listicles meant to sell ad space have seriously harmed the Internet as a whole.

(1)
G_W_Albrecht
Legend
Legend

You have to look at it a different way and it makes sense: A very, very big amount of information, using enormous calculation power, is made accessible by a simple language interface ! I take my head off and make a bow.

I see this as a POC and wait for applications that make more sense 😎

CCSE CCTE CCSM SMB Specialist
the_rock
Legend
Legend

I dont know if anyone here plays chess, but I been playing it since I was 6 or so, so for about 37 years now. Reason I mention chess is because there is strongest chess engine in the world, called Stockfish (I think its up to version 15 now) and its rating is 4200 or so. I mean, to put that in perspective, Magnus Carlsen, who is the current world chess champion, has rating of about almost 2900 and I read somewhere it would take him at least 20 years to get to 4200 and he would have to win 90% of his games to reach that number.

Now, in my personal opinion, and Im sure lots of other chess players would tell you the same, chess engines defy literally all the logic and beauty of the game as people had known for such a long time. I prefer to enjoy the beauty of it the way great Paul Morphy, Adolf Andersen, Bobby Fisher and other great ones played the game.

So, to summarize, yes, engines are more and more present in our lives, but I just find it that it takes away from the normal life people used to live, thats all.

Just my personal opinion.

_Val_
Admin
Admin

@the_rock this is a perfect definition of cyber-phobia, thanks for the example. I am pretty sure those British luddites from the past also have all kinds of great, very human, and very convincing arguments against those evil machines taking soles out of things 🙂

the_rock
Legend
Legend

My only phobia in life is frogs (hate those things lol). Other than that, like 99% of chess players, we think Stockfish is a scumbag : - )

0 Kudos
Tony_Graham
Advisor

From a legal standpoint you would have to assign legal rights to AI, much

in the way Picard argued Data as a sentient being.

Vladimir
Champion
Champion

I think they have just solved the math issues, at least I am seeing it im my Discord notifications. But in all seriousness, the problem with it is the training cutoff date of 2021 and the occasional erroneous answers.

Despite mandatory disclaimers, people tend to copy/paste the code it generates. So, I would not trust it too much.

It'll be interesting to include CP CLI and API in its next training set to see if it'll be able to generate complex queries based on natural language requests then.

Leaderboard

Epsum factorial non deposit quid pro quo hic escorol.

Upcoming Events

    CheckMates Events