- someone responds with a Gen AI-enabled service that is
1) really impressive
2) is currently offered pretty much for free
3) doesn't have that many tangible benefits.
There's many technologies for which it's very easy to answer "how does it improve life of an average person": the desktop, the internet, the iPhone. I don't think Udio is anything like these. Long-term, how profitable do you expect a Udio-like application to be? Who would pay money to use this service?
It's just hard to imagine how you can turn this technology into a valuable product. Which isn't to say it's impossible: gen-AI is definitely quite capable and people are learning how to integrate it into products that can turn a profit. But @futureshock's point was that it is the AI investment bubble that's losing hype, and I think that's inevitable: people are realizing there are many issues with a technology that is super impressive but hard to productize.
why is it that any mention of Rust on the internet will bring about these wildly-overshooting claims. The amount of production software having been re-written in Rust (or replaced by a Rust counterpart) is minimal. This kind of rewrite is rare for any programming language. Yet claims like this are everywhere.
I find it annoying either way: it doesn't provide valuable/novel information and isn't conducive to productive conversation (in part because it's unclear if it's sarcastic or earnest)
which is why it's not a good analogy: you make a powerful tool like "programming language" easier to use, you get more people with the power to write programs. You make an unreliable and brittle tool like ChatGPT easier to use, you get more crap content on the Internet.
Not saying that there won't be good use-cases for Large Language Models, just saying that the analogy doesn't work: these things will be useful if used by competent people in very specific use-cases. The way ChatGPT is used now is just to generate a lot of hype and middlebrow content: with that in mind, I think its low barrier of entry is actually a negative (it causes things like this: https://www.npr.org/2023/02/24/1159286436/ai-chatbot-chatgpt...)
The ratio of dumb stuff to reasonable stuff being said about ChatGPT is dizzying; I truly hope it has reached its peak and this kind of hype and poorly-thought-out analysis dies down soon.
This whole article could be summarized as "How to protect your career? Be good at your job!"
The 'software dev' section is particularly off-putting. It starts off with a code-snippet that is ridiculously wrong: it defines 'typos' as any string matching this regex: '(\w+)\s+\w+'.
That's right, if your text contains a word followed by whitespace followed by a word, that's a typo.
The code could only appear correct to someone with no knowledge of the domain and no experience coding. The fact that this was published on "wearedevelopers.com" is a bit sad.
The article then goes on to deliver this piece of wisdom: "Coding will soon become no more than a means to an end. Which it always was."
OK, if it always was, then why is this news, or a disruptive change? Also, is anybody involved with writing production software convinced that coding is somehow not a means to an end? What else would it be?
I don't have much knowledge of the other careers mentioned in the article, but I suspect the analysis and advice there is about as flawed. The suggestion that teachers should have ChatGPT grade papers while they spend more time "face-to-face with struggling students" seems particularly vapid and shortsighted as well. How would they even know who is really struggling if they're not reading their work, delegating it to a notoriously unreliable tool?
>> The 'software dev' section is particularly off-putting. It starts off with a code-snippet that is ridiculously wrong: it defines 'typos' as any string matching this regex: '(\w+)\s+\w+'. That's right, if your text contains a word followed by whitespace followed by a word, that's a typo. The code could only appear correct to someone with no knowledge of the domain and no experience coding. The fact that this was published on "wearedevelopers.com" is a bit sad.
If you want this hype cycle to come to its inevitable end sooner rather than later, maybe it's a good idea to keep quiet about such errors.
Anyway I'm still curious to see what happens when enough of that garbage code makes it into github for the next iteration of Codex or some similar model, to be trained (or fine-tuned anyway). If we keep pointing out the errors in code generated by LLMs, we risk never finding out.
I'm with you, but I also think that discussion about specific tools often is wrongly prioritized over discussion of general principles. The lesson should be "always profile your programs, and be especially wary of third-party dependencies" rather than "avoid this specific library at all costs."
I'm assuming they also understandably don't want to start a flame war, where their specific criticism of the library is going to be taken to mean "this library is always bad, do not use it, even if its effect on performance is negligible in your case."
Many in the US tend to associate Christianity with right-wing ideology, but there's actually quite a few examples of a varied history of left-wing Christian thought and action. Sticking to US history: Reverend Martin Luther King, Jimmy Carter, Dorothy Day (https://en.wikipedia.org/wiki/Dorothy_Day).
Outside of the US, Liberation Theology in Latin America was an ideology/movement explicitly linking Christian (mostly Catholic) teachings with Marxist-inspired struggles for workers' rights (sometimes those links are easy to do, if you focus on quotes such as "blessed are the poor").
I do agree that FOS software is not socialism, and that donations aren't socialism. But it's important to note that many people who work on FOS software do so without expecting anything in return. I think that idea is in sharp contrast with typical right-wing and centrist economic theories that view human beings as mostly money-driven, self-centered individuals.
We're falling in the pitfall of ‶left and right are different depending on where you're standing″. I'm 100% sure what you say is true in your country; but in mine, being religious and eschewing the state-based, tax-backed wealth redistribution in favor of the personal, charity-based one definitely puts you as left-wing.
Christianity is fully compatible with a society where people can make meaningful decisions over things which affect them, with democratic control their workplace and community. In fact Tolstoy came to this conclusion, that Christian ideology says we should have no bosses, and not dominate or control other people.
I suppose there are various right-wing ideologies, and some of them may be compatible with Christianity. I cannot imagine any left-wing ideology (which is at best wrong and at worst downright evil) compatible with Christianity.
What is mysterious about Chekhov’s success? Genuinely curious as someone who has only read a few of his stories but finds them clearly very well written and influential. Maybe the extent of his success is still somewhat surprising?
I am curious too but the article is in the style of an English major and I just can't stand that style of writing. It is like the literary equivalent to guitar shredders.
I know he was a doctor and that is obviously an odd career trajectory but I don't know much beyond that.
>I don't see what it is about the Twitter product that would require going "hardcore".
It's been more than twenty years since the dot-com bust and people still think that just because a company is on a website that makes it a tech company. Sure, Twitter's product is twitter.com the website, but most criticism of it and ideas for improvement don't revolve around the engineering aspects of keeping a website running: they revolve around 'product' features and sociological ideas on how to organize the community of users, what kind of content to allow/promote and what to censor.
Likewise, when people talk about the value they see in Twitter, nobody talks about the great distributed system that runs behind the scenes, because its only role is to provide users with the actual product: people go on Twitter to keep up with the news, see funny posts, feel connected with other people during major events or while watching a TV show. The tech is only relevant in that it allows this to happen. By that logic Walmart is a tech company because I can buy stuff on their website.
Thinking that by having engineers go "hardcore" you're automatically going to have a better product is like if a pizza delivery company decided to invest in better cars because people criticized their pizzas.
Sure, if you want to build new features fast you'll need software developers working hard, but much more importantly, you'll need people who understand the product and the users (in their various different communities and habits). That expertise is much harder to quantify and it seems clear that Musk deems it irrelevant, at least judging by the rollout of his most recent features and the utter failure of Twitter Blue
> Thinking that by having engineers go "hardcore" you're automatically going to have a better product is like if a pizza delivery company decided to invest in better cars because people criticized their pizzas.
Bingo. Hard-core engineering has been the lynchpin of Tesla and SpaceX's success -- they literally won via designing and manufacturing superior physical things.
It makes sense that EM would blunder by attempting to reproduce that type of success at Twitter. And it will be a blunder, unless: (a) he plans to build one or more new products (or drastically change the nature of the existing product), and (b) those new or drastically changed products will somehow "win" in a big way that Twitter isn't winning now.
But at least to me, (a) doesn't seem likely and (b) seems even less likely.
(a) seems very likely to me. Didn't he already speak about making Twitter a "super-app"? super-apps are a bit new to the American audience but in China / India they are surprisingly commonplace.
I won't be surprised if he churns through a bunch of different ideas: Subscriptions, Deals with news companies, short form videos + creator tools, different kinds of ad experiences, the edit button, better spam filtering etc.
There's a lot of potential things Twitter could be doing if they didn't limit themselves to their core product.
I don't think (b) is unlikely. Their current product is already highly valuable. You don't need their short form video product to succeed in a way TikTok did but there's enough of an audience to capture there to improve the company's valuation.
Maybe I'm living in a filter bubble, but I think it's going to be hard to build a "super app" that relies on the presence and active participation of reputable brands--publishers, if it's a news platform; businesses, if it's a b2c communications platform--when those brands are currently freaked out by the seemingly erratic behavior of the new CEO.
Even if Musk somehow learns to stop live-tweeting his trial-by-fire, the damage might be done: natural iterations on business plans and features will play into a by now near-universal stereotype of Musk as some sort of shoot-from-the-hip madman, which will deter business partners from committing.
(Like, if I'm Fox News or the NY Times, do I want to commit to publish on Musk's new platform when tomorrow he might get stoned and decide to offer hosting to the Daily Stormer? Or, if I'm Bank of America, do I want to commit to consumer payments when tomorrow Musk might pivot his "super app" to offering video shorts instead?)
Again, maybe I'm living in a bubble, but Musk's reputation seems like a huge impediment to, like, doing this. At all. Ever.
He has. But doesn’t the concept of a super app fly in direct opposition with some of his stated goals and issues regarding Twitter? He has ranted about bloat, both in terms of features, engineering, and employment at Twitter, and all his seemingly disastrous moves since taking over have been aimed at reducing bloat. And yet creating a “super app” sounds like it will be adding back in bloat orders of magnitude larger than whatever currently exists at Twitter.
I’m not sure anything Elon says about the current or future state of Twitter has any credibility. It seems like the company is now being directed via stream of consciousness at this point.
Good point you raise. Without broaching speculative argument as to the general state of Mr Musk’s consciousness-stream:
Giant, bloated organizations (and projects, and other things) can often be helped by reduced-size, highly-directed streams of consciousness suddenly making unilateral changes to them.
One way to term it— and there are many— is crisis restructuring. Whether Twitter was actually in crisis, and whether the unilateral stream of consciousness doing the restructuring is ”divinely inspired,” remains to be seen.
Twitter's cost was based on its market value. Expanding the portfolio, however, isn't as expensive, especially not when your new boss has all the money. That combined with its pull factors is why some people bet on Twitter coming around eventually.
On the other hand, exactly that development towards super-apps makes it even more important that we push for decentralization.
Not to mention that the US is already reigned by mono- and oligopolies in many markets, it is also a privacy disaster.
Many nations, especially in Asia, have little awareness of the concept of privacy. and thus its consequences, so it isn't surprising such offers took off so swiftly there.
You are correct, he did mention (a) as his intent in buying Twitter. Makes me think that he acquired Twitter for the users and wants to introduce those super-app features ASAP.
That may be true for its fit and finish (panel alignments etc.), but the fact that they mastered vertical integration and supply chain management in less than 10 years is beyond amazing. Their powertrain (battery, motors, other hv components like charger etc) are first class and unparalleled in performance (and cost). We’ll see what Lucid will achieve in the next couple of years, their claimed (and validated) numbers such as range are very impressive, but for now, but Tesla has proven that it is a hardcore engineering company in a few important areas, there are others that need to catch up.
Tesla's manufacturing process engineering is unparalleled in the automotive space, but the final products are often buggy because when you reinvent manufacturing overnight and don't nail it 100% all people see are the bugs.
I personally don't like Teslas that much, I think they got a lot of things wrong, but the manufacturing process engineering really is leading the whole car industry by quite a ways.
Check out MunroLive's teardown videos on YouTube where they talk about the changes to part counts, castings, fasteners, and all the internal details that consumers don't notice.
Are you familiar with the innovation diffusion curve[0]? It is very normal for industry leaders to build buggy products, because early adopters value innovation more than perfection. That’s how the whole tech industry works.
Tesla may or may not succeed in appealing to the late majority and laggards, but if they don’t, the Toyota you buy in 15 years will still have benefitted from the industry changes Tesla is leading today.
What is Tesla doing in 2022 that every other car maker hasn't started doing? It seems like they're still building buggy products long after this excuse could work.
Well, its hard when you say things like "every other car maker". Other car makers are starting to do some of the same things, but sometimes these are things Tesla did first.
For example, both Lucid and Rivian use round cells for batteries. BMW is exploring the 4680 form factor. Most other manufacturers still do not use round cells.
Tesla is using large castings for major parts of the frame. I think Volvo is starting to do this? Its not entirely a new trend, but again, Tesla was the first in the auto industry to do it this way.
The way Tesla does the structural pack for the 4680 is unique as well at the moment.
Having said that, I think Tesla people overstate the customer significance of all of these things. The real impact is technically interesting, but isn't really making a big difference in the outcome of the vehicles.
There are other EVs that charge faster and are significantly cheaper too, now that Tesla keeps raising prices.
If you find yourself characterizing something as an "excuse", it's a pretty good sign that you're looking for confirmation of bias rather than new information to form an opinion about.
But sure, how about:
- Model years. Have you seen the 2022 Tesla Model 3? No, because it doesn't exist. Teslas evolve over time. Other manufacturers still have days where the next model year starts being produced and the cars are very different than the previous day.
- Greatly simplified cabin. I hate this, but I think it's the future.
- OTA software updates. Some other makers are doing small updates a bit, but nobody (let alone everybody) is doing what Tesla is here
>Teslas evolve over time. Other manufacturers still have days where the next model year starts being produced and the cars are very different than the previous day
People who like Tesla and people who hate it both seem prone to making declarations about how different it is that seem unfounded to me.
The auto industry underwent a long evolution to regimented model years, and they still make changes in between to this day. Framing a lack of model years as progress rather than regress by ~a half century seems arbitrary to me.
When cars are recalled, it isn't uncommon for the population to be defined as a range of serial numbers. The fact that recalls aren't always based on model years seems like pretty good evidence that changes are made in between, doesn't it?
I thought the exact same thing, there's really absolutely nothing special about Tesla anymore.
I appreciate the direction they pushed the industry towards, but I feel like the innovation has finished and I'd expect a lot of the problems to be ironed out.
Toyota on the other hand have produced a car which runs on hydrogen, it is in production and for sale in the USA right now. They also do hybrids and I'm sure if they wanted to, they'd do electric no problems.
American automotive manufacturing became hyper-conservative over time. A willingness to let (non-safety-critical (1)) bugs get to market with the tradeoff of making products you otherwise wouldn't is a market-distinguishing tradeoff.
(1) But my larger concern is that I can't say about Tesla that only non-safety-critical bugs are making it out the door, so I also won't buy one. Because unlike a buggy phone or buggy smart watch, a buggy car with safety-critical bugs can kill other people, so there's wisdom in disallowing it.
Which doubles the ironic nature because most of the Japanese cars sold in the US are built in the US, and many (most) of the "US Cars" that are sold in the US are built in Mexico and Canada.
Why... unions. Japanese Manufacturers have evaded the UAW problem, where Ford and GM have had to flee the nation to get out from under decades of sunk costs
It still absolutely rocks if acceleration speed is your most important metric (which I personally find quite compelling), but in pretty much every other aspect they're janky cars compared to similar vehicles in the same price range and even some cheaper ones. Very fun to drive though.
THIS. "Extremely powerful & fast, fun to drive, and the company making 'em is run by an alpha macho guy who also builds space rockets"...yeah, that has huge emotional appeal to a large number of well-to-do males - who would not want to be seen in a Nissan Leaf, Toyota Prius, etc.
Sounds more like cargo-culting boys rather than adult males... I've personally know noone who decides car buys based on this, and this counts also 2 tesla owners. Car sale tax discounts, free charging, plenty of charging spots were the actual reasons for those 2 (myself I am still happily on petrol for next decade at least, thank you)
That's interesting. I don't own a car and I'm not looking for one, but I still had the impression Tesla was leading on range. However, they're beaten at both price levels, potentially with a significant price discount.
Mercedes EQS 450+, 640km, €106,307
Tesla Model S Plaid, 540km, €140,995
Mercedes EQE 350+, 525km, € 79,850
Tesla Model 3 LR DM, 485km, € 60,995
BMW i4 eDrive40, 470km, € 60,630
VW ID.3 Pro S 5Seats 450km, € 43,720
There's a third issue besides price and range in the EV market right now. Wait time. e.g. that VW ID.3 has a wait time of over a year for delivery. That crosses it off my shopping list, unfortunately.
In Europe where I live - VW, BMW and Mercedes do not have a reliable system of recharging for longer trips. Why they are not working to create that is beyond me.
You don't seem to be across what's happening in Europe. All of the companies you mentioned are invested in Ionity which provides 350 kW chargers at over 400 locations:
Europe has standardized on CCS Type 2 Combo for charging. Any CCS EV can charge on any CCS charging network. Teslas can charge on Ionity, BMWs can charge on Tesla chargers, charging networks like FastNed, BP, GridServe, EnBW, Circle K and friends can charge all brands, etc.:
It is still not that great in the USA. In the Pacific Northwest, if I want to goto Spokane, I have to hope the the Electrify America charging stations in Ellensburg aren't down again. One point of failure doesn't make me feel very comfortable, especially since EA is so flaky (if Tesla opens up there supercharger network, I'll be the first to buy an adapter).
If I want to go down to John Day national monument in eastern Oregon, things are even worse. Things will get better (oh, and we want to do a trip to Anchorage someday...).
That is one of the best tests out there, IMO. I do wish they'd add their charge curve details to it somehow. Something like the estimated time to completely a 600 mile trip would be an interesting detail, as range isn't the only thing that matters.
Aren't all EV ranges measured based on the latest EU measurement standard? Which means cars are incredibly prepped for those tests, but the results are not made up and they are more or less comparable between models and brands.
make sure to researched tested vs claimed range, Teslas have good range but fall short from claimed range by large %, there are cars that end up having the same or better range like the Taycan even though the claimed range from Porsche is less! do your own research, correlation is not your coin, etc
I can't agree with that. They are certainly not janky compared to the Leaf or Bolt and up until fairly recently the 3 was price competitive with the Bolt.
Beyond that, comparisons get more complicated, but there are still perks to the 3/Y. For example, the Ioniq 5 and EV6 charge faster but have less real world range at highway speeds. In practice, this negates most of their charging advantage.
If you want the best EV sedan for road tripping in the US right now, there is nothing lower price and also better than the 3 LR.
Only if we look at the only competition of Tesla being other electric cars, which has never been the case. It was just its most distinguishing feature, but ultimately it's an implementation detail.
If you want to argue that there is really nothing superior about Tesla's engineering that's fine, but you can't shove implementation details under the rug when it suits you. For there to be nothing superior in their engineering there must also be no implementation detail that is superior.
My "implementation detail" comment above does not refer to the quality of engineering overall, but specifically in the context of competition. Yes, when comparing electric cars, the quality of the electric engine is a very large element; but it is much smaller if we're comparing all the elements that matter to someone buying a car (quality of body construction, interior finishing, steering etc).
Why should they bother? They can go into electrics at their own pace, when the infrastructure and demand is better, and still eat Tesla's lunch.
Heck, even Audi A.G. that you've mentioned (not close to being the biggest car company) is comparable to Tesla numbers-wise.
In many industries it's not the first movers that get the market, it's the big mature market-friendly solutions (sort of how the iPod wasn't the first commercial mp3 player).
It's not like people want them to optimize for acceleration anyway...
>Why should they bother? They can go into electrics at their own pace, when the infrastructure and demand is better, and still eat Tesla's lunch.
I had a lecture about electric mobility last semester where a tech lead (don't want to dox them) from Daimler's electric truck program and a tech lead from Daimler's electric car program were invited. In regards to electric trucks things seem to be going great but the electric car program had massive problems with Tesla. The person in question was very hyped about the competition but admitted they got beat with the previous gen cars and that they had to completely redesign their processes to be quicker since Tesla is constantly updating their models. They were hoping to be slightly better than Tesla this gen and to get ahead next gen. So in summary I'd say that didn't turn out great for them.
>It's not like people want them to optimize for acceleration anyway...
The many YouTube views of teslas beating sports cars would beg to differ. If people didn't want acceleration everyone would be driving an 80HP hatchback or a 120-160HP SUV.
So, being on par (on even a tad ahead) of the biggest competitor in the current gen, and being ahead of the competition in the next gen, is bad? Sounds fairly good to me, especially when you have trucks, ICE and EV, and ICE cars that provide a lot of cash to finance all of that development.
That's kind of the issue. Daimler's EV car project is considered a flop because it's only alive due to Daimler being able to lose money from it.
If they started earlier they would have been able to earn some of that money that went to Tesla, could have ironed out the issues while EVs were still considered in something like an open beta where owners had a lot more tolerance for issues and their processes would have become more agile naturally due to the competition with Tesla. Not to mention they could have improved their image as a luxury EV manufacturer by being one of the first.
You can even see the panic about being late to the game on Daimler's part. Their first gen EVs was made using ICE chassis which logically could not compete with cars designed from the ground up to be electric.
They can't enter a new market at their own pace when it is eating demand for their existing market. If competitors wait another 10 years to bring a serious EV to market, they won't have the revenue to cover their expenses on their existing ICE business lines, much less fund the capex needed to get into EVs. And most car companies have a lot of debt, so the problem will be even worse.
Right now EVs make up such a minimal percentage of the car market that they aren't making a meaningful dent in ICE demand, but once they hit 10% or more, car companies that don't have a slice of that segment are going to be feeling it financially. I expect there will be major casualties in the car business because of failure to get into the EV market fast enough.
Benz is a little behind, but Porsche and Audi are doing just fine, in general. The Taycan and etron gt are both fantastic vehicles that compare well with the Model S. The larger etron SUV is about to get a revision that will make it very competitive as well.
The only one that I dislike within Audi is the Q4 etron. Its too close to the id4, IMO.
Tesla showed that there is a viable market for electric cars if the look and work like "regular" cars, not tiny spaceships, and if the supporting infrastructure is in place (Superchargers).
Tesla was not the first EV sold in the US. Earlier attempts by the established manufacturers were curiosities at best. And now that Tesla proved that EVs can actually be good, and can be made profitable, those established manufacturers are willing to move into that market.
So, no, it's not about engineering, it's about taking the risk and committing to expanding the market.
Exactly, Tesla also pushed the envelope with charging speeds, at a time when a lot of people in the industry were limiting things to 50kw and treating EVs as little purely city cars.
The iMiev preceded the Tesla and sold better in many part of the world because it was a more decent proposal.
What Tesla did was making greenwashing and wasting energy with electricity dfast and cool. It is probably better than wasting with ICE in nuclear powered countries but that is not what the world need right now or ever.
Not sure about engineering.
But they build a electric-car factory from scratch in Berlin in 2 years during the pandemic.
Check how many years other companies take to build ICE or electric car factories…
Not sure how “completed” it is, but it already produced Model Y, right?
Compare that with VW, how many electric cars they are making, and how long did the biggest auto-manufacter in Germany (not a bad country for Engineering standards) took to get factories ready for EVs?
Not compared to manufacturers with decades of experience no. But making cars is hard. Starting Tesla from a clean sheet is impressive.
Think about all the Chinese manufacturers, that for years were churning out the butts of jokes. How many of those would have survived in an actual competitive market?
No. But he took it from the proverbial garden shed to a global company.
I'm not a musk fan boy, but what he did with Tesla is impressive.
Compared to the company they were buying chassis from (lotus). If someone took lotus to the size of Tesla in less than 2 decades, would you not say that's impressive?
Starting a boutique sports car manufacturer is 'relatively' easy. Making a global car company, not so much.
Yeah, his skill is in hucksterism. That can be used for good (Tesla), but it can also be used for ill (all the hyperloop nonsense). But he's still a huckster.
When I use the term "hard-core engineering", I'm not referring to how "hard" people work. I'm referring to how advanced the required engineering is to accomplish a goal.
I think this is a really interesting move, that, if Elon hadn't fucked things up so badly, could actually have a chance of making the acquisition work. I think that TikTok is a political issue and the government would like to (and IMO should) kill it because of the Chinese issue, but won't because they fear the user backlash. If Vine could gain enough of a user base to be a credible fallback option, then banning TikTok from the US market would be politically feasible, and Vine would be worth a fortune.
I do not think twitter has the ML capabilities you need to built a TT competitor, and they aren’t just competing with TT anymore. YouTube, IG, and FB are already getting pretty close. Elon would have to something pretty different than just copying them
SpaceX did some good standard mainstream solutions. Sometimes success is not about some exotic innovation but doing your normal job well. A standard lox kerosene gas generator rocket engine. Flat aluminum panel construction. Only the interstage is composite. Started small with Falcon 1. The Merlin engine had heritage from NASA's FASTRAC program.
If you look at Ariane V, Atlas V or Delta IV, they all have some peculiarities and operational warts. Large solid rockets, Russian engine, hydrogen. In some sense, Falcon 9 is the most "ordinary" of the launchers available now.
He also spent 44 billion on it, which he had, instead of using said 44 billion to lobby politicians of both parties and invest directly into whatever he wanted to do with the weapons program money. Sounds like a really solid strategy to me.
Lobbying is surprisingly cheap. The entire pharma industry spends $5B/year, while the entire fossil fuel industry spends $2B/year[0]. Both have achieved a large degree of control over government decisions affecting them.
Elon could have had his weapons program for $1B. He could have dominated defense spending for a decade for $25B. I don’t think the ROI of this $44B of very indirect investment will compare.
Keep in mind he actually spent around half that. There were significant loans and other investments contributing.
He needs more than lobbying, he needs to sway an election. But there were obviously other reasons as well, he probably thinks a hive mind is cool to own and he could finally build his everything app X, etc.. it's just the orbital weapons program is near the top of that list as he thinks it's necessary to save humanity from itself. From free-wheeling philosophy discussions, he seems to think this SkyNet system is actually a kind of Roko's Basilisk.
It was $13B in loans from a variety of banks. Lenders makes yes/no decision on loans, they do not retain management interest the way equity investors do.
This is what I hate about politics and political reporting. People can make a vaguely plausible sounding insinuation about anything being connected. And then noone can disprove it.
But.. isn't he trying to turn it into a Super App like WeChat? That means they have to build about 10 other huge products plus every new idea the boss comes out with in by next week. After having their headcount cut in half. Sounds pretty hardcore to me
If he’s trying to do that, then I’d say he should focus outside of engineering and more on product design, marketing, and brand.
“How to build a Super App” is not that complicated, from an engineering perspective. It does requires hard work, especially if you want to build it fast.
But the big problem is not really how to build it as much as how to make it successful: how to make people use your app over other ones they already use. There are apps like Venmo and WhatsApp that cover WeChat’s use cases pretty well, why should people switch to Twitter?
Step one would probably be to get people to trust your app. I don’t think the latest developments at Twitter make it seem particularly trustworthy
> the great distributed system that runs behind the scenes
As someone is reported to have been asked in a Twitter job interview, when a celebrity with 10,000,000 followers sends out a tweet, how do you deliver them all within 3 seconds?
Did someone measure the 3 seconds delay on tweets sent to 10 mil followers ? Or just some assumption based on lower scale tests ?
I'd imagine "shard the servers serving the subscription to users, accessing sharded database/queue of new tweets" would get you close. Much easier to do with established software now than when twitter was created tho.
if my memory serves me correctly, for "standard" users, tweets are push - it will be pushed to all your subscribers, so they get instant feeds. for "celebrities", tweets are pull - when you open your feed it will pull the latest tweets for your celebrities from a dedicated cache for celebrity tweets
Who in their right mind would rely on Tweets for a calamity??
Phones have built-in support for emergency messages (that bypass all normal silences and notification hiders, unless you explicitly go into settings and disable emergency message support), and carriers are required by law to broadcast such messages to everyone in range.
Twitter is a niche social network (at least in most of the world) and even of those that use it, a huge number will have various forms of disabled notifications.
I personally am rarely on Twitter but it used to be an invaluable resource when you're in an area and don't really have time for a fully fleshed out news article to come out the next day about what's going on on the ground. Block club Chicago was tweeting and retweeting videos of police beating up protesters that night a couple blocks from my apartment in 2020, well ahead of the articles that came out the next day.
It was also interesting to see in real time an active disinformation campaign materialize and disappear that same day. Up until around 1 AM there were a lot of people posting about how the "Proud Boys" were planning to show up or had shown up and encouraging people to bring weapons, then all initial tweets were deleted late at night, with only screen shots or people who had quoted them still being available on the site. But that was a problem I hope would be fixed instead of the removal of the checks to make it all in to a shit show.
Sure, I'm not saying Twitter had no value or that Musk isn't destroying what value it had - he clearly is.
I'm just pointing out a limited fact - that the 3s limit is not relevant even for emergency situations, as no one does or should rely on emergency tweets.
It's used for notifying residents of earthquake, tsunami, and other disaster risks. Residents in japan for most immediate disasters will hear an alert through the national system (via phones, TV, radio, etc) and on top of that residents can check the NERV twitter or other sources to determine what degree of threat it is and what the next best course of action is.
Was like "Huh, didn't know Anno took the name of a real organisation for his show", but from seeing the logo is the logo of the fictional anime organisation, the inspiration is clearly reversed. It seems a real app that gives notifications from actual earthquake warning systems, I wonder if they licensed that logo.
EDIT: Apparently they did some sort of cross promo deal with 3.0+1.0 after existing for a few years, so Khara is aware of them so they must have some sort of agreement to use that name and logo
Agreed. That's what the J-alert system is for. It sends out your basic "get to cover", "stay covered", "find high ground" type alerts but the NERV bulletins give you quite a bit more useful information on where else is affected and what impact that will have. Also worth noting that NERV supports a mobile app which can give you only the useful subset of notifications.
IMHO J-alert is for the first second and NERV is for the seconds after.
I mean, their original setup during early years (and I'm not making this up) required them to restart their Rails service every day, else it would crash/stuck because of memory leaks.
That's not a "tech startup" based on "hardcore engineering", that's a startup based on a social idea and implemented with some trivial tech.
Scaling this to hundreds of millions is not exactly a "hardcore engineering" thing either. Mostly regular engineering with some good practices (like share nothing) and devops. Tons of companies do just the same, today you can even do it from your bedroom by building on top of AWS, GAE, etc, there are even templates for this...
I think you're dramatically underestimating the scope and scale of the problems Twitter had to solve. See the thread below for examples of the sort of solutions they had to build because open source or cloud-hosted solutions weren't available or viable.
>Do you think they still run Ruby that can't reclaim allocated unmanaged memory?
No, I don't.
My point is their initial success years (which wasn't that off the current product) wasn't some hardcore engineering feat: just a cruder-than-average Rails app.
Around 2012-2015 we had someone on-call over the holidays do a redeploy every few days because of Rails memory growth at our now-IPO'd ecommerce giant.
This growth impacted every rails app back then, the popular ones more quickly is all..
It is not about engineering, it is about conformity. As you implicitly acknowledge, Twitter does not need as much engineering, nor does it need creative engineering. For engineering, Elon needs cogs, preferably those who buy into his cult of personality.
As for the hard to quantify work, Elon needs people he can trust. Even if the existing people are good, they can not be trusted. Committing to something like Twitter 2 is as good an oath as it gets. The other alternative is to just fire everyone he did not hire himself.
Engineering, for most digital products, is as good as much you can ignore it exists.
I remember the dread that must be dealing with their process load when I get to a whale page, not when I'm tweeting successfully.
Anyway, the thing here is that since EM wants to enact that much change in a short time frame as possible, and guessing he's likely to make mistakes along the way, he'll need engineering work as hard as possible.
> It's been more than twenty years since the dot-com bust and people still think that just because a company is on a website that makes it a tech company.
Yes like Instagram and others. It was not tech stack that was sold but community of users. I guess Musk's lenders knows this.
> company is on a website that makes it a tech company
Well, we generally consider social media companies tech companies. And since they don't really have anything else other than the tech stack (and the users on it, but they don't "have" those), not sure as what you would characterise it.
> ideas for improvement don't revolve around the engineering aspects of keeping a website running
What makes you think that the desired engineering improvements are about "keeping [the] website running"?
It seems to me that what is being sought are engineering solutions to, for example, stop the spread misinformation without censorship. That would be pretty cool, I think. And I can imagine a way of pulling it off, because we actually have such mechanisms in the (non-internet) social networks that make up what Jonathan Rauch calls The Constitution of Knowledge.[1]
These mechanisms use network effects to amplify good information and attenuate (but not censor) bad information. What is good and bad? The network decides this by what information it attenuates or amplifies. Bad scientific papers don't get cited, good ones do. The stories in the National Enquirer don't get picked up by other papers, and nobody believes them to be true.
This obviously isn't perfect, but it's about as good as it's gonna get, and works pretty well in practice.
Internet social media disrupted these mechanisms by its extreme virality, which seems to apply specifically to bad information. It just spreads extremely fast, literally exponentially (not the "very large" meaning of "exponential"). Kind of like a virus.
And we learned recently that to suppress a viral pandemic, completely 100% sterilising immunity is not actually needed. You just need to slow it down enough to get the R number below 1, ideally significantly below 1. That's what I think the "Do you want to read the article first before retweeting?" prompts are for: slow things down a bit.
Of course "slowing things down a bit" is counter to your economic self-interest and fiduciary responsibility when you're an ad-driven company. So getting off ads completely or at least partly is an important part of it.
> Thinking that by having engineers go "hardcore" you're automatically going to have a better product
Who thinks that? Why do you believe that anyone thinks this? While I don't agree with Elon's approach, the causation would clearly be the other way around: there is a lot of engineering to do to get to a better product (less dependent on ads, more attenuation of bad information without censorship etc.). And all that has to be done while under a bit of pressure, this ain't a Greenfield project with unlimited runway. So yeah, might get a bit "hardcore".
> Well, we generally consider social media companies tech companies. And since they don't really have anything else other than the tech stack (and the users on it, but they don't "have" those), not sure as what you would characterise it.
Consider where the bulk of the revenue comes from and you’ll find out what industry the company is actually in.
So ads.. as much as I dislike the industry, that's also an eng intensive task, right, as you attempt to predict user tendencies, raise engagement by showing the right shit, capture more data, and engage in the bot arms race.
Interestingly enough, it turns out that Twitter never really built a DR system for performance ads, hence why the brands pulling out can crater their revenue.
nope, it's direct response advertising, which is the stuff that gets you to click and purchase online.
Both FB and Google have really successful systems for this, and it insulates them massively from brand safety concerns. It appears that Twitter never invested in this (which is kinda insane tbh), and hence the large brand advertisers have a lot more power over them.
Maybe youre right about them pursuing engagement for its own sake, maybe they're not in that game.
Though, whatever approach you have in discovery, or ranking posts, you're having an impact on your users, and promote a certain dynamic, no? I imagine their next approach won't be the equivalent of letting go of the steering wheel, just maybe not mindlessly chasing the same metrics as they are currently.
Maybe that's a little too social psych to call tech/engineering, but I imagine the data science that would support that would be quite exciting.
Facebook and Google have huge tech operations, but Facebook and Twitter are content distributors and market aggregators, not tech companies.
You can buy a jet or a rocket, but you can't buy a Twitter or a Facebook. You can't even hire part of their stack. (FB has content deals and APIs for advertisers and marketers, but not - so far as I know - direct access to the servers.)
Which is why AWS is a tech company. But Amazon is an online store that happens to use tech.
It's like saying book publishers are really just printers, or FedEx is really an airline and truck driving company.
Some publishers do indeed do their R&D for print operations and logistics. But they're still primarily content houses. Any engineering that happens is a means to a productive end, not a product in itself.
I always thought tech was used as the lever in a tech company..
Doesn't matter what you sell, or service you offer, but if you're leveraging technology to add large numbers of users with a marginal increase in cost, I see that as a tech company.
This generally means there is an internal focus on the technology itself as any improvement can have a direct link to a user's LTV.
But maybe I'm talking less tech company, and more tech led organisation..
> engineering solutions to, for example, stop the spread misinformation without censorship.
I really don't think this is an engineering problem. If I was a tech lead and a product manager came up to me with this, I'd tell them they need to figure out what they mean before they start talking to engineers. Software-Development is not concerned with the nature of truth and misinformation, it is concerned with what works given precise goals.
Nobody in engineering school or at an engineering job gets taught or learns what counts as misinformation and how to assess it: you need people with humanities expertise for this.
The "Do you want to read..." prompts are actually a good example of what I mean, because the engineering that goes into making those is quite trivial. But the idea of using them, trying to make them effective at their goal, and determining how much they're working, these are all things that fall outside of engineering, and have more to do with product-design and sociology.
Creating a feature is not as important as getting people to use it, and use it the right way.
Based on the features Elon has rolled out so far (Twitter Blue), it seems to me they're putting more emphasis on just building it, rather than figuring out how people will use it and whether it will be beneficial
By that definition no tech company is a tech company. People search things on google because they want to find information , not because of the algorithms that bring that information. People use stable diffusion because they want to create images , whatever the cool ai technology behind it.
I would argue that when the primary products of a company are technologies, they are tech companies. That would include TSMC, AMD, AWS and much of Microsoft.
If your products are just BUILT using technology, well then you may consider yourself in another business. Whether that is a clothing, cars, pharmaceutical, weaponry or social media platforms (sorted by increasing potential for harm ;)
Still, even if you're in the last category, it may be fair to consider it "tech" if the main challenges/differentiating factors in creating the product is scienc/engineering/technological rather than for instance design, marketing or organizational efficiency.
Exactly. In the real world no one cares what tech stack you use, whether it's a monolith or not. Businesses serve customer needs. Tech is the tool and is only useful to that extent.
If uses the point is, if you ask a typical WeChat user to describe what it is, you'll get a similar answer to the android equivalent. From a user point of view, they're comparable.
"I have investigated a number of
cases like this, and in every one the data has been correctly recorded in the
database, but for some reason it is displayed/exported incorrectly."
- someone claims that Gen AI is overhyped
- someone responds with a Gen AI-enabled service that is
There's many technologies for which it's very easy to answer "how does it improve life of an average person": the desktop, the internet, the iPhone. I don't think Udio is anything like these. Long-term, how profitable do you expect a Udio-like application to be? Who would pay money to use this service?It's just hard to imagine how you can turn this technology into a valuable product. Which isn't to say it's impossible: gen-AI is definitely quite capable and people are learning how to integrate it into products that can turn a profit. But @futureshock's point was that it is the AI investment bubble that's losing hype, and I think that's inevitable: people are realizing there are many issues with a technology that is super impressive but hard to productize.