BRN Discussion Ongoing

[M_C]

Yan Lacunn seems to be singing a different tune these days........ thoughts??

YC: No, it does, for this model for autonomous AI that I've proposed, which has a single configurable world model simulation engine for the purpose of planning and imagining the future and filling in the blanks of things that you cannot completely observe. There is a computational advantage to having a single model that is configurable. Having a single-engine that you configure may allow the system to share that knowledge across tasks, things that are common to everything in the world that you've learned by observation or things like basic logic. It's much more efficient to have that big model that you configure than to have a completely separate model for different tasks which may have to be trained separately. But we see this already, right?
.
It used to be, back in the old days at Facebook -- when it was still called Facebook, the vision we were using for analyzing images, to do ranking and filtering, we had specialized neural nets, specialized convolutional nets, basically, for different tasks. And now we have one gigantic one that does everything. We used to have half a dozen ConvNets; now, we have only one.

So, we see that convergence. We even have architectures now that do everything: they do vision, they do text, they do speech, with a single architecture. They have to be trained separately for the three tasks, but this work, data2vec, it's a self-supervised approach.

ZDNet: Most intriguing! Thank you for your time.
.
.

Meta's AI luminary LeCun explores deep learning's energy frontier | ZDNet

www-zdnet-com.cdn.ampproject.org

 

[mrgds]

Hi Esq
I just realised you left off I am having a hair cut tomorrow afternoon. FF

So, does this mean FF, you will now be referred to as our new CEO? ............ as in, Mr Shaun Hair ? .........

 

[Fact Finder]

That is so bad it is hairlerious.

 

[Fact Finder]

By the way I highly recommend reading @uiux Brainchip BH thread from the wee hours of this morning. The dots @uiux has put together make an interesting read.

Amazing piece of work. Thanks @uiux.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Stable Genius]

@FU BACON (formerly mc) and others, what is Brainchips relationship with Luminar? I can’t find a direct link and it isn’t on our ‘wall’. This could by quite exciting!

Hi DHM,

I recall looking at Luminar a while back and Renasas were a supplier/partner of sorts so we could end up providing our IP via them.

 

[Stable Genius]

Here’s a cool video featuring Luminar and where they’re headed with adas.

Luminar

www.luminartech.com

 

[Fact Finder]

I just read the following paragraph in an otherwise grey analyst market report:

“Over the last 7 days, the US market has risen 2.7%, driven by gains in every sector and the Information Technology sector leading the way. As for the past year, the market is still up 9.2%”

There is a thing called the rule of 72. It works like this if you divide 72 by the rate of return (9.2%) it will tell you how long it will take to double your money. So at the present rate money invested in the US market will double in value in 7.8 years.

The other day someone posted that Brainchip over the last 12 months at its present levels is up about 30% and on the rule of 72 at this rate of return you will double your investment in 2.4 years. Three times faster than you would in the US Markets.

I am not being insensitive to those who bought at say $1.50 but in judging this company indeed any company I think it is sensible as we approach its AGM to look at where it was at the AGM 12 months ago and what has happened in the intervening period.

Presently it has increased shareholder value by 30%.

It has put in place or achieved every goal it set at the AGM.

They have announced licence and customer engagements and grown the business across every front.

They have achieved and exceeded every technology target they set excluding the release of the USB Stick though in fairness they did not give a precise date for its release.

They have shown increased income and clearly are making sales of Raspberry Pi etc;

They have greatly increased the companies profile and finally buried the opportunity for manipulators to question the validity of their technology.

In short on paper and in reality they have had a very solid 12 months.

Even if you do not accept that the present share price is a product of the Puto effect you would have to accord some responsibility for the present oversold position to him and as such the 12 month return on investment should probably be much higher than 30% at this moment in time.

If it was 50% then you would double the value of your investment in just 1.44 years.

My opinion and thoughts DYOR
FF

AKIDA BALLISTA

 

[wilzy123]

That is so bad it is hairlerious.

 

[TechGirl]

Here’s a cool video featuring Luminar and where they’re headed with adas.

Luminar

www.luminartech.com

Haha that was an awesome video, they are basically paying out on Tesla in it & saying are you ready for real autonomy, awesome, thanks SG

 

[buena suerte :-)]

Here’s a cool video featuring Luminar and where they’re headed with adas.

Luminar

www.luminartech.com

Enjoyed watching that .. thanks SG

 

[mrgds]

Here’s a cool video featuring Luminar and where they’re headed with adas.

Luminar

www.luminartech.com

Mmmmm, .......... now where have i heard the terms, .................. were in a "sweet spot for Lidar"
.................. BRN tech will become "ubiquitous"

 

[Bravo]

So, does this mean FF, you will now be referred to as our new CEO? ............ as in, Mr Shaun Hair ? .........

Or Shorn Hair even...

 

[Sam]

Look who I found on the LTR thread on HC

 

[Bravo]

Check out these revenues! They are astronomical! To think we only need to get our mitts on just a smidge of this and we'd all be rather jolly I'd say.

Samsung retains top spot over Intel in semiconductor market

Samsung has maintained its position atop the semiconductor industry, while Intel has just held its top spot as the leader in annual revenues.

itbrief.com.au

 

[Chilling]

Look who I found on the LTR thread on HC

Seems fair definitely a time to be taking profits soon if you have been lucky enough to enjoy the meteoric rises on some of them.
The lithium price will return to fair value soon enough and most aren't producing any product at all along way off.
Put some profits into semiconductors such as BRN

 

[Fact Finder]

Look who I found on the LTR thread on HC

We need an indifferent emoji. FF

 

[Dhm]

https://www.luminartech.com/mb/

Not sure if posted before but underlines the link with Luminar and Merx.

 

[TechGirl]

I was wondering about the Mercedes Luminar Lidar partnership that was announced in Jan this year & what it meant for the Valeo Mercedes Brainchip connection we have. I came across this article that shows how Luminar in using Valeo sensors for its Lidar.

Mercedes teams with Luminar on lidar sensor technology

Mercedes-Benz and lidar expert Luminar Technologies, Inc. have entered into a partnership to develop future technologies for highly automated driving.

www.eenewsautomotive.com

Business news | January 23, 2022

Mercedes-Benz and lidar expert Luminar Technologies, Inc. have entered into a partnership to develop future technologies for highly automated driving.

By Christoph Hammerschmidt

---

Share:

---

The partnership with Luminar is intended to underline Mercedes-Benz’s approach to expanding its network of high-calibre technology partners and suppliers. The aim of the collaboration with Luminar is to further develop the US company’s current lidar technology so that these lidar sensors can be integrated into Mercedes-Benz’s vehicle series production. The aim is also to shorten development cycles and thus ensure the availability of the latest technologies for Mercedes-Benz vehicles.

“Luminar fits perfectly into our existing collaborations with leading and innovative technology companies,” says Markus Schäfer, CTO of Mercedes-Benz AG. “Mercedes-Benz has already reached an important milestone by achieving the SAE Level 3 standard in automated driving. Collaborations are a key part of our strategy.”

As part of this partnership, Mercedes-Benz Luminar will enable continuous product improvements and updates with a data-driven development approach.

Just recently, Mercedes-Benz became the first automotive manufacturer in the world to receive an internationally valid system approval for highly automated driving (SAE Level 3) – a milestone in automotive development. The “Drive Pilot” highly automated driving system is to go into series production this year in the S-Class and the EQS. However, the lidar sensors for these models do not come from Luminar. Instead, Valeo supplies these components. Luminar, on the other hand, supplies its technology not only to Mercedes Benz, but also to Daimler Truck AG and Volvo. In addition, Luminar cooperates with Nvidia and Mobileye – so one can assume that their platforms for autonomous driving are equipped with Luminar lidars.

http://www.mercedes-benz.com/
www.luminartech.com

 

[Slymeat]

Where are you seeing the publication date?

I thought the publication date was at the top right of the page. But I now notice that has been updated to today's date so it may just be the date the page is feed to the viewer.

There is a list of news items below the article which is up-to-date - this gave me the impression the article was current.

I have attempted to send an email to the identified contact - but that has now bounced.

 

[Lex555]

These type of dilemmas are the reason I had thought I would never see fully autonomous vehicles on public roads (at least not autonomous vehicles controlled by a Von Neumann based computer), yet they exist.

The situations mentioned by @Jumpchooks are but a couple of situations that CAN be thought of and programmed before-hand. That is the trouble with a Von Neumann based system. It needs a set of instructions and set of data that is coded up BEFORE-HAND. This only works with deterministic problems - those for which everything is known, at least for the solution space that is desired.

The trouble I see with a fully autonomous vehicle is that the problem is non-deterministic. I.e. it contains things that are unknown and even things that can never be known before hand. Even the sheer number of knowns is immense.

And as you say @Jumpchooks, the human psyche has a way of dealing with unknowns on-the-fly—intuition, instinct, luck, guess work and pre-emotive thinking are some of the ways of handling unknowns. I suppose this is why neuromorphic processors where first conceived; emulating brain neurons and synapses in a hope to emulate the way the brain forms new pathways to handle new situations, and the more those pathways are exercised the more permanent they become. Hence learning and adaptability to change—a mechanism to deal with unknowns. Isn’t it even more amazing that Brainchip have been able to actually produce this in silicon!

So for the non-deterministic problem of a fully autonomous vehicle we are left with a “good enough” solution. And if that “good enough” solution is better than the best driver on the road, that may even be the “best solution” or at least an improvement and a benefit to humanity. I’m actually ok with that!

How do we determine the “better than” in the above. Basically this is being determined by accident avoidance and responsiveness. But once again, this is ONLY for known situations.

Statistics are gathered that show less accidents, so that is deemed positive.

But statistics also prove more serious accidents as autonomous vehicles tend to plough into perceived non-threatening obstacles at full speed. Crash investigations have shown that fatal autonomous vehicle accidents tend to not have had any attempt to apply the brakes. Turning what could have been a less serious accident into a fatal one.

Remember the Tesla that slammed into a table-top truck turning in front of it on the freeway. Due to heading toward the sun that was low on the horizon, the car was somewhat blinded and did not see the truck. The Tesla hence carried on at speed, resulting in the death of its occupant. An astute human driver may have slowed down as the sunlight conditions made it difficult to fully appreciate what lay ahead on the road. A human driver most likely would have thought to shade the sun. The Tesla had not been programmed to make such a judgement call.

There’s also the dilemma of do least harm and at what cost. When are the occupants of the vehicle deemed less worthy of saving than a group of pedestrians, for instance? Is it based solely on numbers? Or do the autonomous vehicle designers program to protect the vehicle occupants at all costs? That probably is how a human in command of a vehicle would think!

A human driver can say I didn‘t have time to think. But a computer program will most likely always have ample time to run through a large number of potential outcomes and pick one that best meats its programmed parameters. And whoever designs those parameters should be accountable for the outcome.

This adds strength to the recent Mercedes announcement that the company will take responsibility for accidents. I fail to see how this will actually apply in the case of a fatality, and especially internationally. Will some company exec or developer wear man slaughter, or even murder, charges if proven that the design/coding had malicious intent either deliberately or even by omission?

I suppose only time will tell.

It will all come down to the value of a life metric, multiplied by the expected number of lives lost compared to the number of lives saved (assuming that is possible to determine). It will all come down to money!

Personally I would never be a willing passenger in a driverless vehicle. I like the act of driving and of being in control. Getting from point a to point b is not the sole purpose of any trip I go on. I derive a lot of pleasure out of the act of driving.

I am 100% supportive of driver assist technology as it can enable the driver to enjoy more of the driving experience by taking care of some of the more mundane tasks (e.g. sticking to the speed limit) and it does improve safety.

I sympathise with your love of driving but give AI progress over say 10-20 years, and it’ll be hard for society to justify a human driver who by then will all be labeled incompetent, to use the same road way with AI. The varaiation between man and machine by then will be akin to justifying drink driving.

Human driving, along with combustion driving will probably be a very popular race track experience