Wednesday, December 4th 2019
Intel in Negotiations for Habana Labs Acquisition
Intel is currently performing negotiations to acquire Israeli AI chip startup, Habana Labs, according to a person who spoke to Calcalist anonymously. If the deal realizes, Intel will pay between one and two billion USD, making it Intel's second-largest acquisition of an Israeli company. When asked about the potential deal, the Intel spokesperson has stated that the company will not respond to rumors surrounding it, while Habana Labs has yet to respond to a request for comment made by Calcalist.
Founded in 2016 by Israeli entrepreneur Avigdor Willenz, who founded Galileo Technologies and Annapurna Labs, Habana Labs develops processors for training and inference of Machine Learning models. This acquisition would allow Intel to compete better in the AI processor market and get new customers which were previously exclusive to Habana Labs.
Source:
Calcalist
Founded in 2016 by Israeli entrepreneur Avigdor Willenz, who founded Galileo Technologies and Annapurna Labs, Habana Labs develops processors for training and inference of Machine Learning models. This acquisition would allow Intel to compete better in the AI processor market and get new customers which were previously exclusive to Habana Labs.
18 Comments on Intel in Negotiations for Habana Labs Acquisition
Intel's clearly heading the way of Cyberdyne :shadedshu:
So yeah, retail demand for AI accelerators is rather small, but so is for bare Zen cores or RAM dies.
AI chips are already arriving inside SoCs and appliances. You'll have many of them at home soon enough.Yes, exactly. Today you let your washing machine decide how much water it needs. Tomorrow it'll grab a knife and stab you.
Come to think of it, it never was a priority I guess.
But this problem is in a different dimension anyway. Intel is not moving away from CPUs and into AI. It's more about enhancing processors so that they do AI tasks faster. Because suddenly AI is not about complex problems in science and engineering. It's for basic stuff as well.
Washing machine programmed in the factory will never be very efficient. It's made for an average client: with average water quality, using average detergent and loading average cloths.
You give that washing machine some computing power and sensors - it'll configure itself much better for the particular bathroom it'll work in (kitchen in US, I guess). And it'll keep learning with each cycle.
This will be a general trend in coming years.
But a washing machine that can get angry with how dirty your house is, so it decides to flood it... that's at least decades away. Don't worry too much. :)
en.wikipedia.org/wiki/Artificial_consciousness
I been wondering what would the motives be for a washing machine to kill the owner. Dirty house might not be right. The robot house keeper's to blame :P Humans are save :D
logo will do just that - 'grab a knife and stab you'.
Thanks a lot guys, look what you're going to start. Intel machine uprising.
Desktops are already a thing of the past - at least outside of gaming.
They're kept by DIY enthusiasts, so component makers try to monetize on this (just like car makers still making V8 engines and Leica still in business making 60-years-old camera bodies).
Intel focused on mobile and servers over 10 years ago.
CPUs are a different story. Intel is not going to sacrifice anything. In fact they keep growing.That's because for an AI it's easier to judge and punish than to feel empathy etc. So it will try to kill you if this maximizes a utility function (i.e. it thinks your death is good for it or for community).
Think about people with road rage. They see someone forcing the right of way or parking on grass and they'll say all kind of stuff: kill him, throw a rock, call the police. Most of the time they won't do anything. A machine would. :)
But yeah... for a machine to be able to kill you or flood your house, you'd have to implement an instruction that does it. So don't. :)
Self aware computers that can make complex decisions and expand their code when needed... that's a distant future. But it's 100% possible.
I don't think Intel is out of desktop market yet. (gaming doesn't envelop entire desktop market actually it is not even worth mentioning here) There's still a lot of sales for Intel in that department. Inter is slowly moving away from it. AI and accelerators for concrete purpose is What Intel is after.Surprised you went forward with this. As a sarcasm and a joke, it didn't have to be commented or explained pointing out machines don't have feelings but I get it :)
You say: Intel is slowly moving away from desktops and they're after AI and accelerators.
And what about the rest of the processor market? The "larger rest" to be exact. :)
As I said: they're not focusing on desktops for 10-15 years already.
It's a market they're making a lot of money from, but they're not going to spend on long-time investments, because these just wouldn't pay off.
Mobile, servers, IoT - these are the hardware markets worth investing in.
"AI" is just a umbrella-term for some technology. It's not a product. It'll be in everything (desktops as well), but it's not something Intel can sell as a hardware maker.I'm not sure why that would be seen as a joke. It's in our future.
Computers will be able to feel emotions at some point (or do something extremely close), but you don't need emotions to kill someone. Deciding that it's beneficial is enough. AI cars are being programmed for this already.
Don't get cocky with the phrase "it's in our future". Nobody knows what's in store for humankind and if it's going to happen it's in far, far future. I doubt we have a decade for AI to rise against humankind. Don't watch sci-fi too much.
BTW to kill someone you can't have emotions, be emotionless and empty inside, not that you don't need emotions. You don't need emotions to cook a meal or mow the lawn for instance.