Fully automated AI networks less than 5 years away, reckons Juniper CEO

AI will completely automate the network within five years, Juniper CEO Rami Rahim boasted during the company’s Global Summit this week. “I truly believe that just as there is this need today for a self-driving automobile, the future is around a self-driving network where humans literally have to do nothing,” he said. “It's …

[Neil Barnes]

Lies, damn lies,

and statistics?

[The Man Who Fell To Earth]

Re: Lies, damn lies,

I hope it works better that Tesla's "Autopilot" or we're all doomed.

[amanfromMars 1]

Get with the ProgramMing..... FFS

Too much like shades of ...“It was him/her/them/IT/an AI what done it, your honour, not us working with Intel” .... to be very different and anything novel ...... https://forums.theregister.com/forum/all/2022/05/12/intel_amd_security/#c_4459118

[An_Old_Dog]

Debugging

Hmm, and how do we puny-minded humans fix an AI-driven network when it breaks? What's the mean-time-to-repair on that? Or is AI (which was itself created by puny-minded, error-prone humans) somehow magically-infallible?

[amanfromMars 1]

Re: Debugging

They don’t break An_Old_Dog and don’t break, An_Old_Dog. No worries. Bugs are for entomologists.

[Zack Mollusc]

Re: Debugging

Duh. you just tell it to fix itself!

[hoola]

Re: Debugging

Ahh but that is the point....

If it is AI then it must be perfect and never go wrong.

Of course this totally ignores the issue that it has been created by humans who have proved notoriously bad at producing software that it no bug-ridden

AI is just software and a set of rules.

But it makes great sales waffle and reasons to screw money out of companies.

[veti]

Re: Debugging

The AI fixes it, obviously.

If the AI breaks down, reboot it.

How exactly is that different to having a human network manager? The AI will be quicker to respond, and probably better at explaining its decisions.

[AndrueC]

An interesting idea.

[Anonymous Coward]

Isn't that 'interesting idea' basically 'SKYNET' of Terminator fame !!! ???

:)

[Ken Moorhouse]

2 + 2 = 5

Repeat it often enough and it will become the norm.

If everything ever recorded throughout history is the truth then AI might arguably stand a chance of converging on reality. But this is not the case. Anyone that says that Black is White, or 2 + 2 = 5, distorts that convergence. Yes it is possible to "prove" that 2 + 2 = 5, but only if an intermediate result which makes use of a questionable use of zero is glossed over. The "artificial" in AI arguably aims to mimic "intelligence" and, as we know from history, even intelligent people get things wrong from time to time.

Specific to this particular assertion about automating networks: If a packet is routed through a network a billion times, 99.999% of the time it might be routed one way. The nature of intelligence means that the rest of the time it may be routed a different way. This is the way we test some facts, by positing a fallacy as truth, then knocking that assertion down. While traversing that erroneous trail, statistics will reinforce that error as truth, so it will take a very large network indeed to converge those results. The state of the chip market at the present time is arguably unlikely to feed such over-indulgence.

[Snowy]

Re: 2 + 2 = 5

2+2 is five for the correct value of 2 :)

[SCP]

Re: 2 + 2 = 5

I thought it was:

2+2=5 for sufficiently large values of 2.

[SCP]

Re: 2 + 2 = 5

Yes it is possible to "prove" that 2 + 2 = 5, but only if an intermediate result which makes use of a questionable use of zero is glossed over.

The proof of that I saw involved a sneaky disregarding of a negative square root. Very sweet (but I do enjoy false proofs)!

[Plest]

Re: 2 + 2 = 5

"Anyone that says that Black is White..."

Could get killed on the next Zebra Crossing!

[ecofeco]

LOL wut?

Bwhahahahahaha. Go on, piss on my other leg.

xkcd: Dependency

[Sixtiesplastictrektableware]

robot kids off my LAN

With all sincere gratitude, thanks for that.

A good little laugh goes such a long way these days.

Thanks again.

[pimppetgaeghsr]

Will the next market crash bring these clearly false AI predictions back down to the trough of disillusionment

There is so much money being dumped into these niche AI startups where young engineers correlate model performance to something that can actually generate money in the future.

What a waste.

[amanfromMars 1]

What a waste indeed, compounded by the fact that no lessons are learned, thus guaranteeing failure.

There is so much money being dumped into these niche AI startups where young engineers correlate model performance to something that can actually generate money in the future.

What a waste. ..... pimppetgaeghsr

Any such sums of money pale into insignificance whenever one considers the wasted performance of all the vast sums received and squandered via the means and manner of compulsory government taxation on practically anything and everything thought taxable but which itself generates zero money for the future.

And such is why all current well established markets are clearly crashing or in imminent danger of crushing collapse and flash crash.

[Plest]

Yeah sure...

2022 and state of AI....may be able tell the difference between a cat and a dog if fed 150,000 clear images of both.

Larry Ellison, twat that he is, predicated cloudy type stuff would be the norm back in the 1990, well it only took a mere 25 years and we finally started renting PCs on the internet to run Windows/Linux insead of running them in cold rooms under the office space!

AI can only be about another 50 years away, so I'm not going to lose any sleep.