4 ms·

> [Andy Jassey] said something like this: “All these leaders [CIOs and CTOs of huge enterprises] are asking me what our blockchain strategy is. They tell me tha

> [Andy Jassey] said something like this: “All these leaders [CIOs and CTOs of huge enterprises] are asking me what our blockchain strategy is. They tell me that everyone’s saying it’s the future, the platform that’s going to obsolete everything else. I need to have a good answer for them. I’ll be honest, when they explain why it’s wonderful I just don’t get it. You guys got to go figure it out for us.” To me the tell is not just that Andy didn't understand. It's that all of these leaders said "everyone says it's the future", but not one of them said "I have this problem and here's how blockchain solves it for me."

TrackerFF 4y ago
I recently attended a trade show, which mostly revolved around emerging ML/AI and other "hot" tech products in the market. So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). Some researchers working on these products repeated that they often get approached by directors and CxO level people regarding this - wondering if blockchain can "solve problems". And every time, they have to reply: Using the blockchain can solve some specialized problems, but there's no magic fairy dust involved, and it's not some general structure which will solve all their problems. Unfortunately, it's kind of like AI. High-level directors dream of General AI, but you need to explain to them that it's all ML models which aid (replace if you're lucky) workers, and automate some tasks. I think any CTO worth their salt should know these things, and manage the expectations of the other C-suite executives.
- masklinn 4y ago
  > So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). Literally none of that should involve (let alone requires) a blockchain.
  - TrackerFF 4y ago
    I'm no expert on that, but we had a bunch of senior researchers (NGO) advocate blockchain for that exact task. I'm not going to argue against people with combined decades and decades of research and development experience in the food logistics community.
    - foobiekr 4y ago
      Such people latch onto trends to look smart. Then they scrub history, as most investors are doing wrt blockchain, when trends change.
      - masklinn 4y ago
        > Such people latch onto trends to look smart. Or because they figure the hype will allow them to finally sell (launder) the project they need to management. smt88's top-level comment (https://news.ycombinator.com/item?id=33688576 https://news.ycombinator.com/item?id=33688576) is a thing I've seen multiple people mention.
    - lowbloodsugar 4y ago
      Using the definition of blockchain as: "A blockchain is a type of distributed ledger technology (DLT) that consists of growing list of records, called blocks, that are securely linked together using cryptography." Then, no, you don't need "blockchain" for that. You don't need it to be distributed. You don't need them to be linked using cryptography. In fact there are good reasons not to use blockchain, such as "I don't want my competitors seeing the exact details of my business transactions". I've heard, "oh but we can fix that with encryption", and then silence when asked "so how are you going to distribute those keys?" etc. Looked into this for large medical supply chain company. Blockchain isn't merely no use, it's actively unfit for purpose.
- throwoutway 4y ago
  they’ve been talking about that agtech solution for over 5 years and no one has built it? They could do the same with a database
  - jasonwatkinspdx 4y ago
    A friend of mine is an economist that did a similar project (sans blockchain) in grad school. He set up a program to label fish with information about where it was caught, who caught it, when, how it was processed etc. Some of it on the package including a smiling photo of the boat owner, more details on the web behind a QR code/URL. He worked with a local chain roughly equivalent to Whole foods and had a cold case full of product like this in several of their stores. People seemed to like it when asked, but when given the choice to pay a premium for it vs the unlabeled fish, the majority declined. And this is at a store that caters to the whole hippies with money crowd. I think the people hawking this idea for consumers utterly fail to take into account that most consumers simply do not care about this. A non trivial portion actually do NOT want to know this information, similar to how some people are fine eating chicken but can't touch a raw one to cook it.
    - masklinn 4y ago
      > people hawking this idea for consumers Do people actually... do that? I struggle to think of traceability as a massive consumer boon, as far as I could think the primary customers of such would be: 1. states (regulatory agencies) 2. consumer advocacy groups 3. the corporations themselves
- Karunamon 4y ago
  >So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). The usual shitty dismissive HN response to this would be something like "but that doesn't require a block chain it could just be a database". To those people I would ask who is running that database? What are their incentives? Forget about the currency uses of block chains for a moment. The fact that you have this distributed network, controlled by nobody in particular but with their incentives aligned to keep it running and keep it secure, means that you have a platform you can build other stuff on top of. An Ethereum smart contract is a type of append only database, the code of which is strongly open source and very battle tested (basically everyone is using openzeppelin), where you do not have to care about where or by who it is being hosted, with an uptime that puts every hosting provider in existence today to shame.
- Chico75 4y ago
  I'm not sure I see why or how would blockchain solve this information centralization problem better than other "classic" tools?
- diseasedyak 4y ago
  I'm still trying to wrap my head around how a blockchain would be helpful, even in your example. How is tracking say, a cow, from farm to lot to slaughterhouse, and knowing what all was done to said cow, not simply something that could be done easily and fast with a database? Is the portability of the blockchain the thing? I could see maybe how that would help, if there wasn't a standard across the databases involved or something. Honestly just trying to understand, thanks!
  - TrackerFF 4y ago
    The problem with traceability systems comes down to interfacing between the various nodes. Interoperability is probably the largest obstacle on the road to provide a full farm-to-table product tracking. Probably not so much of a problem for the fully vertically integrated companies - especially those that also build or control their electronic systems, but large parts of agriculture still consists of independent actors on each node. IIRC, the argument was that blockchain does not explicitly "solve" the problem, but rather that interoperability between blockchain-based systems will be easier than between those using completely different standards. And of course, then you have the problem that food can travel around the globe many times before ending up on the table. Which again means more trust.
  - coredog64 4y ago
    A database works if you have a central authority that manages it and accepts inputs. I’m going to switch to a different animal for a better perspective. In the Pacific, tuna is caught within the EEZ of a number of (mostly) poor island nations. Some boats are part of the national fleet, but most are foreign flagged. The privilege to fish in an EEZ is costly, so there are a number of measures used to ensure that there is no cheating. The boat owners keep records, there are third party observers on board, and vessels broadcast positions. You need to cross check all these records, and do it in an environment without broadband. To make things more complicated, there are incentives between countries. Fishing days are a finite resource governed by treaties. Poor countries have an incentive to oversell, and it’s possible to look the other way within an EEZ sometimes. Not only that, but countries have legitimate reasons for not broadcasting anything but gross yields from their EEZ. All of this means that there’s not one central authority that can provide a single database.
- UncleMeat 4y ago
  Published transparency records work without blockchains. Certificate Transparency is an example of a project that does this without all the mess of blockchains. And that also works well because it operates on purely digital things - once you start having to unify a physical object with a digital record all the usual supply-chain-management problems re-emerge.
FlacoJones 4y ago
most of the problems blockchain attempts to solve are public coordination problems in adversarial environments rather than corporate problems the CTOs/CIOs are likely to have.
time_to_smile 4y ago
I think one of the hardest parts of becoming an expert in a technical area is the transition from "I don't understand so I must be wrong" to "I don't understand so something must be wrong or I'm not seeing something". When you're more junior you often make the mistake of wrongly dismissing ideas because you think you are very clever. I find that by the time you unlearn this, you typically should reverse direction. Early in my career, every time something didn't make sense, it's because I had a misunderstanding of how things really worked. Much later in my career I started to realize more and more often when I said "this doesn't make sense" I would start assuming I was the one missing something, only to time and time again uncover some one else had messed up or that something else was fundamentally wrong. Obviously even the most expert in their field should reserve some probability that they are misunderstanding, but learning to recognize that you are an expert and that you not understanding something is in fact smoke is an important late career skill.

TrackerFF 4y ago

I recently attended a trade show, which mostly revolved around emerging ML/AI and other "hot" tech products in the market. So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). Some researchers working on these products repeated that they often get approached by directors and CxO level people regarding this - wondering if blockchain can "solve problems". And every time, they have to reply: Using the blockchain can solve some specialized problems, but there's no magic fairy dust involved, and it's not some general structure which will solve all their problems. Unfortunately, it's kind of like AI. High-level directors dream of General AI, but you need to explain to them that it's all ML models which aid (replace if you're lucky) workers, and automate some tasks. I think any CTO worth their salt should know these things, and manage the expectations of the other C-suite executives.

masklinn 4y ago
> So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). Literally none of that should involve (let alone requires) a blockchain.
- TrackerFF 4y ago
  I'm no expert on that, but we had a bunch of senior researchers (NGO) advocate blockchain for that exact task. I'm not going to argue against people with combined decades and decades of research and development experience in the food logistics community.
  - foobiekr 4y ago
    Such people latch onto trends to look smart. Then they scrub history, as most investors are doing wrt blockchain, when trends change.
    - masklinn 4y ago
      > Such people latch onto trends to look smart. Or because they figure the hype will allow them to finally sell (launder) the project they need to management. smt88's top-level comment (https://news.ycombinator.com/item?id=33688576 https://news.ycombinator.com/item?id=33688576) is a thing I've seen multiple people mention.
  - lowbloodsugar 4y ago
    Using the definition of blockchain as: "A blockchain is a type of distributed ledger technology (DLT) that consists of growing list of records, called blocks, that are securely linked together using cryptography." Then, no, you don't need "blockchain" for that. You don't need it to be distributed. You don't need them to be linked using cryptography. In fact there are good reasons not to use blockchain, such as "I don't want my competitors seeing the exact details of my business transactions". I've heard, "oh but we can fix that with encryption", and then silence when asked "so how are you going to distribute those keys?" etc. Looked into this for large medical supply chain company. Blockchain isn't merely no use, it's actively unfit for purpose.
throwoutway 4y ago
they’ve been talking about that agtech solution for over 5 years and no one has built it? They could do the same with a database
- jasonwatkinspdx 4y ago
  A friend of mine is an economist that did a similar project (sans blockchain) in grad school. He set up a program to label fish with information about where it was caught, who caught it, when, how it was processed etc. Some of it on the package including a smiling photo of the boat owner, more details on the web behind a QR code/URL. He worked with a local chain roughly equivalent to Whole foods and had a cold case full of product like this in several of their stores. People seemed to like it when asked, but when given the choice to pay a premium for it vs the unlabeled fish, the majority declined. And this is at a store that caters to the whole hippies with money crowd. I think the people hawking this idea for consumers utterly fail to take into account that most consumers simply do not care about this. A non trivial portion actually do NOT want to know this information, similar to how some people are fine eating chicken but can't touch a raw one to cook it.
  - masklinn 4y ago
    > people hawking this idea for consumers Do people actually... do that? I struggle to think of traceability as a massive consumer boon, as far as I could think the primary customers of such would be: 1. states (regulatory agencies) 2. consumer advocacy groups 3. the corporations themselves
Karunamon 4y ago
>So it actually seems that in the logistics line of things like manufacturing / production / etc. there is a place for blockchain, mainly due to the immutability and decentralized distribution properties. What could take a whole week to track down using older methods, takes seconds if all the data is recorded on the blockchain...for example, food production. You want to find out exactly which animal your food comes from, and all the things involved (use of antibiotics? which factory? how did the product travel before ending up at your grocery store? etc.). The usual shitty dismissive HN response to this would be something like "but that doesn't require a block chain it could just be a database". To those people I would ask who is running that database? What are their incentives? Forget about the currency uses of block chains for a moment. The fact that you have this distributed network, controlled by nobody in particular but with their incentives aligned to keep it running and keep it secure, means that you have a platform you can build other stuff on top of. An Ethereum smart contract is a type of append only database, the code of which is strongly open source and very battle tested (basically everyone is using openzeppelin), where you do not have to care about where or by who it is being hosted, with an uptime that puts every hosting provider in existence today to shame.
Chico75 4y ago
I'm not sure I see why or how would blockchain solve this information centralization problem better than other "classic" tools?
diseasedyak 4y ago
I'm still trying to wrap my head around how a blockchain would be helpful, even in your example. How is tracking say, a cow, from farm to lot to slaughterhouse, and knowing what all was done to said cow, not simply something that could be done easily and fast with a database? Is the portability of the blockchain the thing? I could see maybe how that would help, if there wasn't a standard across the databases involved or something. Honestly just trying to understand, thanks!
- TrackerFF 4y ago
  The problem with traceability systems comes down to interfacing between the various nodes. Interoperability is probably the largest obstacle on the road to provide a full farm-to-table product tracking. Probably not so much of a problem for the fully vertically integrated companies - especially those that also build or control their electronic systems, but large parts of agriculture still consists of independent actors on each node. IIRC, the argument was that blockchain does not explicitly "solve" the problem, but rather that interoperability between blockchain-based systems will be easier than between those using completely different standards. And of course, then you have the problem that food can travel around the globe many times before ending up on the table. Which again means more trust.
- coredog64 4y ago
  A database works if you have a central authority that manages it and accepts inputs. I’m going to switch to a different animal for a better perspective. In the Pacific, tuna is caught within the EEZ of a number of (mostly) poor island nations. Some boats are part of the national fleet, but most are foreign flagged. The privilege to fish in an EEZ is costly, so there are a number of measures used to ensure that there is no cheating. The boat owners keep records, there are third party observers on board, and vessels broadcast positions. You need to cross check all these records, and do it in an environment without broadband. To make things more complicated, there are incentives between countries. Fishing days are a finite resource governed by treaties. Poor countries have an incentive to oversell, and it’s possible to look the other way within an EEZ sometimes. Not only that, but countries have legitimate reasons for not broadcasting anything but gross yields from their EEZ. All of this means that there’s not one central authority that can provide a single database.
UncleMeat 4y ago
Published transparency records work without blockchains. Certificate Transparency is an example of a project that does this without all the mess of blockchains. And that also works well because it operates on purely digital things - once you start having to unify a physical object with a digital record all the usual supply-chain-management problems re-emerge.

FlacoJones 4y ago

most of the problems blockchain attempts to solve are public coordination problems in adversarial environments rather than corporate problems the CTOs/CIOs are likely to have.

time_to_smile 4y ago

I think one of the hardest parts of becoming an expert in a technical area is the transition from "I don't understand so I must be wrong" to "I don't understand so something must be wrong or I'm not seeing something". When you're more junior you often make the mistake of wrongly dismissing ideas because you think you are very clever. I find that by the time you unlearn this, you typically should reverse direction. Early in my career, every time something didn't make sense, it's because I had a misunderstanding of how things really worked. Much later in my career I started to realize more and more often when I said "this doesn't make sense" I would start assuming I was the one missing something, only to time and time again uncover some one else had messed up or that something else was fundamentally wrong. Obviously even the most expert in their field should reserve some probability that they are misunderstanding, but learning to recognize that you are an expert and that you not understanding something is in fact smoke is an important late career skill.