This paper and a recent post by Sebastian Raschka (where he decomposed a Forrester report about the uptake of technologies in industry) is alluding to something I have witnessed in system/control design and applied research.
Both LLMs and massive CV architectures are NOT the holistic solution. Rather, they are the sensors and edge devices that have now improved both the fidelity and reliability to a point where even more interesting things can happen.
I present a relevant use case regarding robotic arm manipulation. Before the latest SOTA CV algorithms were developed, the legacy technology couldn't provide the fidelity and feedback needed. Now, the embedded fusion of control systems, CV models, etc. we are seeing robotic arms that can manipulate and sort items previously deemed to be extremely difficult.
Research appears to follow the same pattern...observations and hypothesis that were once deemed too difficult or impossible at that time to validate are now common (e.g., Einstein's work with relativity).
My head is already spinning on how many companies and non-technical managers/executives are going to be sorely disappointed in the next year or two that Stable Diffusion, Chat GPT, etc. will deliver very little other than massive headaches for the legal, engineering, recruiting teams that will have to deal with this.
Really appreciate articles like these...it somewhat reflects how society expects people to conform, but it's always the outliers that manage to change the world.
The positions of the satellites are well known and documented. Fun fact..on the Garmin fenix watches, it is recommended that you synch your watch with your Garmin Connect app every few days (week) so that you download the latest satellites position file, which significantly improves the GPS lock time with the satellites before you begin an activity.
As others have stated, I don't think Apple's product will really erode Garmin's market share for several years. The lack of detailed topo maps and applications will severely hamper them.
With Apple latest shift, I see their ecosystem as a very strong contender for business travel (international news reporter?) that requires strong security (lockdown mode), ability to communicate in many form factors through adverse conditions (voice, data and now satellite), and has a slew of safety features (crash detection).
Garmin has decades of experience regarding outdoor travel and safety that Apple would always be playing catchup...IMO.
Anyone know performance wise between Garmin's use of Iridium and Globalstar?
Iridium has been around for soo long...I am getting constant outages from the status page. Kind of disconcerting at times.
Ah, I found the following article that somewhat explains the different technologies w/
"The main difference between Iridium and Globalstar is the relaying mechanism. Iridium requires relaying between satellites. Globalstar requires relaying between satellites and earth stations."
I wonder how many ppl will now travel to more remote areas with a false sense of security (and being woefully under prepared) that you can "Just call" for help.
As a paying customer with Garmin's inReach service, I'm acutely aware of how spotty and unreliable the service can be based on environment and current surroundings.
The same amount that did when the cellphone was invented, when Selective Availability was turned off, when the safety match was invented, when the chronometer was invented....
I agree..this article seems very click-baitish, which I sadly clicked on.
What company wouldn't want to hire the best? I've been casually making mental notes of the positions/titles of the groups impacted, and it follows common sense that talent acquisition, administrative, junior level employees, units that no longer support the business mission, and low-performing employees (or ones that don't fit the culture) were let go.
If anything, the article could have pointed out that the performance expectations have been shifted-up a notch or two as the remaining employees (once considered to be high-performers) are now "average". This could become a vicious negative feed-back cycle if not appropriately managed.
sigh Corporate life can definitely be a drag. The holidays can't come fast enough!
I agree. Two very different business responses between Musk and Zuck. I almost feel sorry for Zuck as we are witnessing meltdowns of two large companies in real-time. Musk just continues to dig his own grave…
I feel bad for the ppl losing their jobs but that is a very generous severance package.
Zuck did it himself too. No one told him to go all-in on VR. No one. All he had to do was tackle payment and maybe cloud. He first went for crypto and then is in the process of failing with VR.
I'm not looking to praise Zuck but he did at least try something new. That's how companies stay alive and vital and relevant. Innovation. He gambled big on VR and it didn't pay off. At least not yet. I don't think it will pay off but I have to respect that he went big on a new direction for the company.
I still think Facebook is evil and I feel like they should have tried to buy Tik Tok although I don't know how feasible that ever was.
TikTok is a Chinese company bringing in mountains of data on US citizens, including the ability to influence what people in the US see on a daily basis. The Chinese government would never sell that kind of leverage to anyone, let alone to Facebook, which is banned in China.
TikTok is incredibly popular in the US and there isn’t really a huge similar precedent where the US takes a page from China’s playbook and blocks a social media network completely. If banning TikTok was reasonable, and I’m not sure it is, I think it makes a lot of sense to be cautious about setting bad precedence.
Agree on VR, payments, disagree on Cloud. It's a saturated market, there are half a dozen operators who each have unique selling points. I don't know what Meta's would be.
Doubling down on becoming one of these "everything" apps could have been a good strategy. Become the app frontend for one of the less big food delivery companies in the FB app, tie in to payments. Perhaps even buy Square for Cash App and all the POS integrations to build a network of sellers, all tightly integrated from the consumer perspective into the Facebook app. I'd have hated it, but I suspect it could have worked.
A lot of apps want to become a social network but Facebook was in a good position to do so. Imagine someone posting a picture of their Brunch. They tagged the location. Some AI matches the picture of the food with the pictures from the menu (google reviews already does this to some degree).
That food looks good, imagine if they partnered with Uber or Grab so you can add to cart right below the picture.
Peer to Peer payments could have also been great, especially if you could check-out at a store by scanning a QR code to pay (think WeChat Pay, FairPrice in Singapore, or even Paypal's version of that).
Or even buying event tickets. They already have events on the platform, and they let you put targeted ads, but what about an integrated experience to purchase tickets right on the platform instead of there being an external link?
They could have done so much but the only major change/addition in recent years was Dating (a huge hit in countries that perceive Tinder as only for hookups) and those avatars that people use everywhere instead of text posts.
Very good lesson here for both Twitter, FB and any other upcoming startup. Never treat your 3rd party developers as shit. Look what WeChat achieved with their superapp and developer ecosystem. Twitter and FB tightened their rules a lot over the years, when they had a potential to become super app for West
Interesting. Is that because payments are much lower margin than ads? Surely investors would be smart enough to see the additional revenue, and likely additional benefit to the ads business, as being worth it?
You have to make big bets to continue winning. It's easy for us to sit in our armchairs and criticize their failures, but for example their plays with going mobile-first in 2008 and the acquisition of Instagram in 2012 worked out very well.
It's better to make many small bets and when they start to take off, THEN put the foot on the pedal. Zuck has been notoriously bad at creating new products, so betting the company on that he'll manage it this time seems like a very bad idea.
That's a good point. The principle of proving a simple possible version makes sense. Just pointing out that those "small" steps Tesla took are already a lot bigger than most software projects.
hmm, maybe, but it seems like a golden age for tech where it was hard to fail from a strong starting position. MS, Google, Amazon, Apple, have all done much better than Facebook.
If I put on my rose colored glasses, I still wish Facebook had just stuck to identity.
They could’ve been “the login for the social internet”… they even built that platform! They just were so paranoid about losing control of the graph they shut it all down. Twitter also failed on the developer/platform front for the same reason.
They could’ve been the identity platform for every hot startup in the last 10 years. They could’ve courted developers such that every platform add-on they did got immediate head start… like ads! They could’ve out-AdSensed Adsense.
Anyway, I’m sure that’s all terrible business strategy, but it’s what I wish they had done. Even though I’d probably be cursing their name now if they owned all of our logins.
I thought this problem is less related to vr and more related to ads revenue that dropped because of apple. vr was just a way to create a platform from the ground where ads will continue to be their business model
FB briefly was in the cloud PaaS business when they acquired Parse.
The problem is that the way Meta runs its data centers and software stack is tightly integrated with the products. It’s not really amenable to running third party applications or storing third party data.
Amazon’s infrastructure was also tightly integrated with its products. Despite the often repeated and very wrong myth that AWS was founded by Amazon selling its “excess capacity”, AWS was always created with a separate infrastructure that was purpose built to sell to other companies:
What about the concept of a data center inside a data center? Given their infrastructure size and necessary geographical layout, it should be possible to have a number of IaaS racks stored inside their existing data center footprint.
If they have their own data centers (which I assume they do), this would make a lot of sense, kinda like a ghost kitchen — a virtual data center. That is, assuming they have the physical space to support something like this. It would be a way to diversify income with largely existing resources and vendor contracts.
Imagine even a slimmed down service like fly.io or Cloudflare workers running at FB data center scale.
not a ton of market for that. and it changes the risk nature of their own facilities. already plenty of hyperscale datacenters with space to lease. what advantage does meta offer? surely they wont beat on price.
It’s probably not worth the hassle to FB, but it is funny to think about how big of a business this could potentially be. But even a profitable business unit might not make enough profit to actually make it worthwhile.
It could certainly work. But it would probably be too small a business for a company as large as Meta. The differences in scales is (I think) one of their problems. At Meta scale (somewhat a pun), many things are just harder/not worthwhile because of their size.
SMB. They are effectively the webmaster for a vast amount of very small business, but also Meta's ad platform ends up being one of the larger expenses for many businesses. In fact, I doubt there is a single entity on earth that has more billable B2B relationships.
I agree that spinning up a pure-play public cloud makes no sense for Meta. Its not in their ethos, moreover selling various abstractions over virtualized compute is a commodity. Why would they get in line, behind IBM and Oracle?
Given that Office 365 is being counted as 'Cloud' imagine what Meta could do with some $100/yr SMB service. On the enterprise end, they have some of the very best big data and ML infra and could do well to bundle up extra capacity sell that on a metered basis. If they had started offering managed Presto in 2015 this conversation wouldn't be happening.
Their network infra (IP space, undersea cables, edge pops etc) is also rather vast and I could see a lot of SMB to F500 customers lining up to leverage it if bundled right. If they wanted to they could write a check for CloudFlare, I checked their balance sheet. Meta Cloudflare would be a juggernaut; so powerful that I pray the FTC wouldn't allow it.
Historically Facebook has been allergic to B2B outside of selling ads. Even within it they bought and killed Atlas, effectively handing a monopoly on ad serving to Doubleclick. Now they are warming up to it, offering Workplace, Kustomer, and Oculus for enterprise. I think that the Metaverse could be a novel B2B play and so do they, calling it "The Future of Work".
tl;dr: Meta could win the cloud business because it has the people, cash, differentiated tech, and existing relationships. They could beat AWS/GCP/Azure in many segments of IT spend by packaging their assets together into a novel kind of cloud.
>If they wanted to they could write a check for CloudFlare, I checked their balance sheet. Meta Cloudflare would be a juggernaut; so powerful that I pray the FTC wouldn't allow it.
Why would there be any issue from an FTC standpoint? As far as I can tell, they're in completely separate businesses. I do agree it is a brilliant idea to Microsoft-ize the SMB relationships they already have to sell software services.
I support I feel uncomfortable about it, but maybe such a merger wouldn't raise antitrust flags. CloudFlare has an insane amount centralization. I love their services as a web user, developer, and operator, but WOW do they have a lot of power by nature of their business. I worry about a buyout by a less principled company that could do all manner of wrong with CloudFlare's assets. For example, a Meta Cloudflare could start to delay or block 1.1.1.1 DNS queries to their competition, and do so quietly and selectively. Any service that offers "Protection" ought not be part of a conglomerate.
They don't need to win, they just need to be there as another option. Every business I've worked at has been huge on wanting cloud diversity of some sort, and tons of startups act as middle men on this.
Another of the big boys offering a cloud product would guarantee it would pick up customers and give them another avenue they can plausibly hunt for competitive advantage in.
They should have never given up on mobile. Burning a few billion a year on an android fork would have had much better returns than trying to go warp speed on VR.
Correct me if I am worng, but hasnt Twitter has seen more growth in the last week than it has in some time? 15 million new users isn't a meltdown, nor is thinning a bloated and wasteful enterprise. Also, if twitter goes completely belly up, Musk would still be worth hundreds of billions of dollars. Grave? I'm game for some hyperbole, but not this early.
You being downvoted just shows even this community prefers emotional projections over simple facts.
There is no Twitter meltdown. Before Musk it was already in grave financial trouble and would have made 800m$ cuts anyway. Musk most certainly is clumsy in his actions and communications, but Twitter isn't going anywhere.
Likewise, Facebook isn't having a meltdown either. There's a dent in ad spent against a backdrop of 2 years of dramatic overhiring (same as Google, Stripe).
There's a 4% decline in revenue on a 27 billion quarterly revenue. Meltdown? There's a handful of companies on this planet being this profitable.
Financially twitter is in a rough spot. They were not really making money. Now they have loans to pay too. It isn't in meltdown but certainly there are things to look at there. Sure, Elon can keep it going for as long as he'd like. But he's a fickle personality. I mean he went back and forth several times just with buying the company. Who knows if he'll lose interest.
You're right, but Twitter basically has been in continuous financial trouble since eternity. In 2016 they almost went bankrupt. They tried to sell then but nobody wanted it. Just before Musk they were also in financial catastrophe mode. Under Musk, that will likely continue for at least a year. It's a fundamentally unhealthy businesses.
I'd like to use a common Dutch expression to explain the Twitter situation: "the soup isn't eaten as hot as it is served".
Musk wants absolute free speech but that's just a random interview quote, not the actual plan for Twitter. Users are abandoning the service in droves. No, they are not. A handful of advertisers stop spending (conveniently part of an economic downturn) but that doesn't mean the vast majority do, or do so indefinitely. Twitter is an awful place now, whilst he hasn't implemented a single change yet. Checkmarks will get decimated whilst his original unhinged idea is already dialed back.
Everybody's jumping on all kinds of hysterical projections that are not supported by the facts. There is no meltdown.
Yup, they are worse off in some ways, and better in others. There is a lot of upsides with having a single person like this calling all the shots, and love him or hate him Musk has been successful in the past.
The questions are: can they monetize that, and will it continue? But as far as twitter dying, the opposite is currently true. It's never been more alive.
Huh?. Meta's growth has slowed, but they are a money printing machine. Earnings have gone down because of the enormous bet Zuck has made on the metaverse.
Twitter has been barely scraping by for years. The two companies are not really even comparable.
No, earnings have been going down because Apple fucked Meta with Apple's policy change on asking for consent to be tracked. Meta has even stated as much in various earnings calls since this happened.
This has severely hurt Meta's ad revenue, i.e., earnings.
The metaverse stuff is a bad bet, you are correct, but is not likely impacting earnings in any significant way.
Revenue has slowed from the Apple change, but the drop in Q3 profits can almost entirely be pinned on RL as staff and other investments has accelerated.
> company’s rising costs and expenses, which jumped 19% year over year to $22.1 billion during the quarter.
> Meta’s Reality Labs unit, which is responsible for developing the virtual reality and related augmented reality technology that underpins the yet-to-be built metaverse, has lost $9.4 billion so far in 2022.
The effects from the Apple changes are mostly in the rear view mirror at this point. You could attributed a 4% revenue hit to them, but those can also be attributed to a general slowing economy.
Earnings down 4% in bad macro conditions for advertising and compared to a pandemic when people spent more time on the internet. I don't think facebook revenue is going anywhere soon, a decade from now though who knows.
Their income and operating margin has almost halved, compared to 2021. Their free cash flow is 1/50th of the previous few quarters. Those are truly horrible results.
FB was a money printing machine, but they trashed it.
Everyone wants to dance on Meta's grave, but it's way too soon. Yes, the Apple change gave them a top line haircut, but if the RL spend is excluded, they are making a ton of money. I'd also argue that the real headwinds are the general economy and TikTok.
I don't think Mark saying "we're going to focus on the metaverse instead of making money for a few years" is a meltdown. They're still making money, their earnings are still sky high, they're just spending more than they need to.
Zuck is doing first ever mass layoffs for a company he started from his dorm room 18 years ago and grew to a ~trillion dollar valuation.
Musk is following the standard playbook of private equity takeover + gutting the company to squeeze out remaining profits and then sell for parts. There is no question of even a bit of emotion involved from his side.
Both LLMs and massive CV architectures are NOT the holistic solution. Rather, they are the sensors and edge devices that have now improved both the fidelity and reliability to a point where even more interesting things can happen.
I present a relevant use case regarding robotic arm manipulation. Before the latest SOTA CV algorithms were developed, the legacy technology couldn't provide the fidelity and feedback needed. Now, the embedded fusion of control systems, CV models, etc. we are seeing robotic arms that can manipulate and sort items previously deemed to be extremely difficult.
Research appears to follow the same pattern...observations and hypothesis that were once deemed too difficult or impossible at that time to validate are now common (e.g., Einstein's work with relativity).
My head is already spinning on how many companies and non-technical managers/executives are going to be sorely disappointed in the next year or two that Stable Diffusion, Chat GPT, etc. will deliver very little other than massive headaches for the legal, engineering, recruiting teams that will have to deal with this.