It is very difficult for me to see any amount of money being thrown at Anthropic as a bad idea.
The amount of new revenue that I am personally able to create for my clients, using Claude models for dev, and Claude models inside the insanely agile products delivered, is astounding.
If I was not currently experiencing this myself, and someone told me that this was possible, I would be calling them names.
You could say the same about Codex (and other tooling). Opus as a model is market leading (trading blows with the greatest that OpenAI is peddling), but there will be a reckoning when open weight models are good enough - and I'd argue we are almost there with some of the latest releases. If you hook up the latest OpenAI models to something like OpenCode, its a taste of what an open harness with a powerful model (outside of a providers ecosystem) will be able to offer developers in the future.
> there will be a reckoning when open weight models are good enough
Will you have the hardware to run them? Perhaps. Will enough of Anthropic's/OpenAI's large enterprise customers have the hardware to run them and the money/desire to have their own internal teams set up and maintain them?
You’re paying the subsidized cost. Those margins will shrink once the real bill comes due. I really think everyone will look back at this time as the golden area of cheap AI. We are already seeing the costs (and restrictions/limits) creep up with the Western models.
It amazes me how productive it's possible to be using AI, but I also has this nagging feeling that we are being reeled into being so reliant on this that when the price starts going up, we will simply eat the cost.
The math is pretty simple, and it's easy to justify still paying the price even if it goes up 10 fold, when compared to hirering more resources its still cheap.
So I guess having multiple players and competition in the market is the key?
Going forward, models will start specializing. Anthropic will build a BioMed model for large drug companies. A math/compsci model for frontier theoretical research. A physics modelf or nuclear research. They can communicate each other for synergy effects e.g. for areas where math meets biomed etc. This will be cost reducing as well. We plebs don't need advanced models for our plumbing software work. Following example applied to AI capabilies will make it clear.
Does everyone need a graphing calculator?
Does everyone need a scientific calculator?
Does everyone need a normal calculator?
Does everyone need GeoGebra or Desmos ?
I think the opposite. AI will get cheaper as models become more efficient and we solve the datacenter/energy problem. I bet 10 years from now AI, that is way better than what we have today, will be close to free.
ZIRP and Moore have helped the cloud build up with a promise of profits and ever increasing performance. The future is likely different.
"Power will be cheap" is hope you can hang any hat on. We've been increasing compute per watt but again that's on Moore. I don't think it makes sense to bank on a new energy surplus.
100% agree. I have been trying to tell everyone to build their ideas, and exploit this environment where 100B of VC money into OpenAI/Anthropic = some percentage of money invested into your idea. This is the golden era of building! The music is gonna stop soon. Build now ffs!
> I really think everyone will look back at this time as the golden area of cheap AI.
Chinese models like Deepseek v4 are as good and 10 times cheaper. You can even run Deepseek locally. So no, cheap AI wont be over. Just the US investors won't be able to profit off of the artificial bubble that is there now but wont be in the future.
Would you mind sharing what you can and want about how the sausage is made? I would love to hear concrete cases where actual leverage is measurable. I‘m asking in good faith, not to attack your standpoint.
Same with Codex and very soon with open source & local models. Training great models (for coding and similar tasks) seems to be a question of scale and not much more.
It is likely that 99% of the value created by Anthropic / OpenAI / friends will go the end user. Which is great news.
I believe that I am more of an AI realist. The agentic dev tools are really helping me out, but if I could wave a magic wand to make AI go away for a hundred years, I would do it.
I really hope that we can all laugh at how wrong I was.
However, I believe that the horrors will likely outweigh the benefits. Our global society/political systems are not ready for Stasi as a Service, mass unemployment, or any of this impending crap storm.
To me it is more like software consultant speak than AI booster speak. And it is not exactly surprising that the people in a particular subculture all talk similarly.
Well, I hear it from people who are regular devs and not consultants, although it's more common with people who aren't really working in the trenches anymore.
Like ex-developer turned PM who is now vibe coding everything they can and thinks it's the greatest thing ever.
I've had an account for a while too, and I do think that that GP comment has a style typical of "AI boosters" -- breathless, big on hyperbole, and low on detail.
To the GP: I'd like some details of these "insanely agile products". Is this insane agility reflected by your customers saying that they have a better, faster, more reliable product? How are you measuring this?
It's like insane hype marketing speak because that is genuinely the difference from what it was like to develop software 6 months ago. You see many people using the same language, often in comments that are otherwise stylistically quite different, because many people are experiencing the same thing.
I get that it's tedious to sit on tech forums listening to an endless stream of people insisting that suchandsuch technology is world-changing. Many people and probably most people who say that are wrong. But sometimes the world really does change.
There's substantial observable change pointing towards a universal software development speedup in the neighborhood of 2x. Much of it is internal company metrics, simply because it's meaningless in most enterprise contexts to count how much software is released. Things you can count, like the number of phone apps published, show the same pattern: https://techcrunch.com/2026/04/18/the-app-store-is-booming-a...
I'll grant that there's evidence of more low-level activity, but I'm not sure that equates to meaningful change particularly. "Released an app" is a neutral signal on its own, in much the same way that the Unity asset store led to an increase in game releases, but 'more asset flips' isn't really a major change to the gaming industry.
If software development speed has doubled, then we should be seeing not just an increase in apps being released, but an increase in product output from the big players too.
We should be, but I just don't know how you'd measure that in the short term. It's very hard to put a number to "how much software did Google release this month". You'd expect to see a substantial increase in revenue, but most kinds of software generate revenue only months down the line after adoption picks up, and few companies have even released Q1 results yet.
We've had LLMs for far longer than a quarter now; even if you think that only the recent ones have led to any improvement, you'd still expect to see something.
I started with the prompt "F-104 Starfighter" and it went pretty well. [0]
I wish I could share all of the things I clicked on afterwards, exactly in the order in which I tried to tell the story. When I try to use the share feature on anything below the root node, I get "This page could not be saved for sharing." However, the video generation does work in that share modal, in the way I would expect. [1]
I have been thinking about the parent comment and your reply all day.
Both are exactly my experience, and a really important lived lesson.
I suffer far less, and communicate much more effectively, when I write a work-related email that is crafted specifically for the receiving audience. This often means making it far simpler than I had first imagined.
The human condition when confronted with an email turns out to be: "I ain't reading all that!"
The US Federal Government specifically calls out Monero as one of the coins that it hates, which means that it must be quite effective at achieving its goals. So the pro is that you know it works. The cons are nothing specific to Monero, just general criticism of cryptocurrencies. Not being a deep crypto user myself, at least, I haven't heard anyone speak of any flaw specific to Monero that isn't shared by a significant portion of the remainder of all of these coins
One admittedly minor issue is that since monero is ASIC resistant, its very attractive to run on compromised computers. Other coins suffer this as well I'm sure, but I don't have any data on this. I guess this is mainly a hunch.
> The US Federal Government specifically calls out Monero as one of the coins that it hates, which means that it must be quite effective at achieving its goals. So the pro is that you know it works.
Official condemnation doesn't work like that. Facebook's cryptocurrency, Libra, was also condemned, and we know it didn't work because it never actually existed.
Which is its biggest weakness. Lightning over Bitcoin is decently anonymous, too. Given that is just a layer-2 technology and can be developed further and evolutes outside of Bitcoin protocol changes, makes it more flexible and has shorter innovation cycles.
And outlawing bitcoin has become basically impossible after the large amount of ETF inflows. Monero is nice technology, but I think the ship has already sailed for Bitcoin (and L2 solutions like lightning).
Anonymity is not binary 0 or 1. There is a difference in there is a one-in-a-million chance that you made a payment, is not the same as there is a one-in-three chance you did the payment. You can define anonymity to be a one-in-a-billion chance, and all lower odds as "not anonymous". But that is not applicable in the real world.
Isn´t the experiment in El Salvador proof that Bitcoin does not work as a currency? If you think it isn't, then do you have any measurable pass/fail test for it?
El Salvador is the worst example, as they have been bribed by the IMF to get rid of bitcoin. I would also wan't to learn how the situation in El Salvador would be any different if they adopted Monero instead - I think it would be the same outcode.
I am not a bitcoin-as-currency evangelist. I see it more as digital gold, and gold is not a currency today. It will have its role as a store of value and fallback unit of trade that keeps government currencies in check - another pillar in financial checks-and-balances.
That was not my argument at all. First I said, I don't see bitcoin as currency at all. Second, major nation states have always tryed to use their power to force other nations to use their currency (petrodollar, cough). This is nothing specific to crypto, but something also specific to traditional fiat currencies.
Any high-volatile asset such as bitcoin is IMHO not suited as currency. The good news is, with the bitcoin taproot upgrade and latest lightning standards, you can actually issue stablecoins over bitcoin's taproot asset protocol, and send it over the existing lightning network. My bet is on stablecoins-over-lightning as currency, and bitcoin as store of value. One blockchain to rule them all, other chains not need (for financial transactions at least).
No, Layer-2 systems only transfer cryptographically signed IOUs between nodes.
Settlement only happens when these IOUs are cashed out, and to cash out you need a transaction in the blockchain layer, so the point about latency still stands.
It is as much an IOU as the US Dollar was pre-1971. That is a pretty good image for Lightning/Bitcoins relationship. Lightning is the dollar with a guarantee that you can convert it to gold anytime you like by presenting it at the central bank. Very few people ever converted their USD to the underlying gold as a settlement transaction. The difference with lightning is, the government can't just rug-pull you and stop exchanging those paper bill IOUs - it is cryptograhpically secured that you can always convert to bitcoin. Since no one would consider exchaging dollars as settling in gold, lightning settlement is not tied to on-chain transactions.
Payment channels are possible on other networks as well. Once again, there is no inherent advantage to Bitcoin here. I know because I worked on one (https://raiden.network/). I also dealt with many of its failure modes:
- insufficient liquidity on intemediate nodes
- network partitions
- uncooperative nodes
- nodes that were liquidity sinks and forced other participants to bear the costs of deposits
- insufficient market makers
But more than anything: people do not want to use crypto for payments. It gives them no significant advantage over traditional credit/debit cards, it has no built-in solution for appeals or reversals and it forces them to learn a bunch of stuff to be minimally safe...
> Payment channels are possible on other networks as well
you are moving the goalpost in the discussion of this thread. User KaierPro said bitcoin would not be suitable because transactions takes to long, to which is responded lightning solves that. Now claiming that other cryptos can have layers-2 is correct, but adds nothing to the discussion or my initial point. Yes other chains have faster settlement times, and can have their respective payment channels - no one argued against that.
In practice, it has shown that it is only viable if adoption by number of nodes and TVL grew by orders of magnitude, and both are very unlikely to happen because - like I said - spenders have nothing to gain from it and no matter how much of the UX friction is solved, it will never be as easy as paying with credit card.
The only people who want to use Lightning are the ones who are invested in Bitcoin. Everyone else just want simple/safe access to a payment network.
You are just moving the goalpost again, without adding to the discussion. If spenders have nothing to gain because they prefer creditcards, then this argument applies to bitcoin/lightning, monero and all other cryptos all the same. Nothing to do with my initial point which was comparing lightning/bitcoin to monero.
> If spenders have nothing to gain because they prefer creditcards, then this argument applies to bitcoin/lightning, monero and all other cryptos all the same.
Most spenders will prefer credit cards, but there is a non-zero group where absolute privacy is important and monero is the better choice, therefore more valuable to them.
You are the one trying to make some false equivalency by saying that "bitcoin/lightning is good enough for most cases, therefore there is no need for monero".
The problem is that you are starting with the conclusion that you want (i.e, "Bitcoin is the best") and you are working backwards from this conclusion to make all sorts of rationalizations. Try going from the use case first and then let's see where the reasoning takes you. You will see that for pretty much ANY use-case, Bitcoin is not the answer.
It is splitting words. There is a settlement layer in lightning, which is presenting the preimage and unpeeling the onion HLTCs in reverse order. This happens at the latency of the network path, so usually less than a second. Bitcoin settling is usually tied to confirmation in the block, which lasts ~10minutes. Lightning might be IOUs, but ones that are fungible themselves and not tied to a specific debtor. Actual lightning-to-bitcoin cashout would probably not happen for everyday use, or at least not more often than you change bankaccounts in todays terms.
That's intellectually dishonest. It's like saying wire transfers or card payments are only valid after interbank settlements are finalized.
Bitcoin Lightning is cryptographically designed to be valid even if it's not yet settled on the main layer. It provides cryptographically sound mechanisms to overrule anyone that tries to "cheat". There is no mathematical way to cancel or double spend it, just like your dollars are valid when the transaction is committed in your bank's database although the money still technically hasn't left the other bank.
There are plenty of failure modes where you can lose your funds even if your wallet keys are not compromised. Try running a lightning node, make transactions worth more than a few hundred dollars and then leave your node offline for a few days. Or even more simply: ask yourself what happens to your funds if the disk on your lightning node goes bust and you don't have a recent backup.
THis is the issue, until its settled in the chain, then you are down to trusting the 2nd layer.
Anything offchain has a whole bunch of issues that are either naively or deliberately obscured by the fact that it _eventually_ writes to the blockchain. The exchanges that offer instant settlement are circumventing trust by doing the settlement for you. You get speed, but not security that they have done what they have said they have.
Well, to be fair to OP: small business and retailers are also not getting "real" money when they accept payment via credit cards from Visa/MasterCard.
To be honest I think the issue here is not due to speed of settlement, but layer-2 is not an acceptable substitute because it does not allow reversability. For the merchants it's good that they are getting the money right away, but most consumers will not dare to pay anything via layer-2 networks simply because they won't have any recourse in case they want to undo the transaction.
You can implement reverseability with a credit system, such as Visa/Mastercard. It should not be implemented in base layer or layer-2. It is basically an escrow system.
So now you are proposing to build yet-another piece to an already complex system just so we can justify the existence of your beloved blockchain in the first place.
How long it's going to take you to realize that even if we built everything you are asking for, we are STILL going to end up with a system that is not as capital-efficient as the status quo?
This is just untrue.
If someone cheats in lightning, and you demonstrate they cheated as you describe, then you get all of the locked BTC as a reward. This is on layer 1. Essentially you can easily prove your nonce was signed more recently.
Because lightning uses Bitcoin's blockchain, which is the most secured (as in energy) and the most common (as in market cap) and probably the most accepted as in regulation. Plus, you can use bitcoins taproot asset protocol to issue stablecoins and send them over lightning. No other blockchain needed - which in my opinion, renders monero obsolete or at least a very niche product.
- Lightning is not fully anonymous. It can still be traced by the participating nodes.
- Ethereum's blockchain consumes less energy, is more decentralized, it's a lot more resistent to any form of attack and it can also host fixed-supply coins.
- Market cap is a meaningless metric, it's at best circular logic and at worst it's just a "one billion flies can't be wrong" argumentum ad populum.
- It baffles me how the Bitcoin enthusiasts talk so much about ideology and freedom, but get completely silent about the fact that most mining operations are done in countries with oppressive regimes, financed or subsidized by dictators with access to cheap fossil fuels.
> Lightning is not fully anonymous. It can still be traced by the participating nodes.
"Fully anonymous" is a strong term. Even cash is not fully anonymous. I would give monero that it is more anonymous than lightning because it is a core design principal. There is a spectrum to anonymity, however. As public enemy number one, such as Snowden or BinLaden, your anonymity requirements are different than a citizen buying illegal erectile dysfunction medication online.
If you consider the new features added in lightning over the past 24 months such as trampoline payments, blinded paths etc. - you will find that lightning is anonymous enough. Plus, you can increase anonymity in the client implementation at the expense of higher transaction fees (longer paths, more trampolines). Lightning's BOLT12 standard, which is currently finalized, will increase anonymity even further.
> Ethereum's blockchain consumes less energy, is more decentralized, it's a lot more resistent to any form of attack and it can also host fixed-supply coins.
Thats is factually untrue. First, ethereum famously had a human-coordinated rollback with a controlled restart organized between devs and node runners over Discord. Second, Ethereum is not decentralzied at all, because that is a core property of proof-of-stake: There is no way at any given time that you can be sure that the majority stake is not already in a single entities (or colluding group) possesion - and would thus have absolute control. It is therefore never guaranteed at any given time, that the network is decentralized.
> Market cap is a meaningless metric, it's at best circular logic and at worst it's just a "one billion flies can't be wrong" argumentum ad populum.
Price is ultimately what determines the value of anything. It is absolutely far from meaningless, as the market cap is also a big factor if a crypto asset can be outlawed or banned. Given how many investors in the west already own bitcoin, there would be a massive outcry if it is suddenly outlawed. I say you could outlaw Monero tomorrow and the mainstream media wouldn't even cover it.
> It baffles me how the Bitcoin enthusiasts talk so much about ideology and freedom, but get completely silent about the fact that most mining operations are done in countries with oppressive regimes, financed or subsidized by dictators with access to cheap fossil fuels.
You mean, such as the United States? Because the US (especially Texas) is one of the biggest miners of bitcoin currently.
But you can only make any claims about the properties of a system when looking at the extremes. If Bitcoin's blockchain does not make strong anonymity guarantees as Monero, then Bitcoin by definition can not be the "blockchain to rule them all" that you so desperately want.
>ethereum famously had a human-coordinated rollback with a controlled restart organized between devs and node runners over Discord.
That was achieved through social coordination. No backdoor was exploited, no one had their coins stolen on the original chain. The system worked as intended.
Can you say the same about Bitcoin? Do you think that all these banks and exchanges trading ETFs have secured access to the bitcoins they claim to have? When one of these institutions goes bust, who is going to bail them out?
You keep trying to argue that Bitcoin is more valuable because it is more likely to be supported by the powers-that-be, and that is the strongest indicator that all your evangelism is driven by "Greater Fool" dynamic. Satoshi's idea for crytocurrencies was to have an alternative system that worked despite adversarial governments, yet we keep getting time-and-again evidence that it can only work if it becomes of an instrument for the powerful institutions that caused the problems in the first place.
Bitcoin and its blockchain has no intrinsic value. Unlike Monero, it is not fully anonymous. Unlike Ethereum, it has no utility for decentralized applications. It can not be used as a currency. All Bitcoin has is first-mover advantage and a huge number of people with cognitive dissonance trying to keep the bubble inflated.
> Because the US (especially Texas) is one of the biggest miners of bitcoin currently.
Access to cheap fossil fuels? Check.
Facilitated by the government? Check!
Serving the interests of the elites and the aspirational 14% instead of the general populace? Check!
Really up to you personally what is a pro and a con. For me this is a starting list. A lot of these are a result of the technical differences as well as I listed the technical differences.
Pros:
1% inflation
no fixed supply (makes it more of a currency than an asset)
privacy by default,
fungibility—every coin is the exact same, no coin history
prevents financial surveillance by corporations,
protects against government abuses,
useful tool for activists, journalists, minorities, useful for domestic abuse survivors,
useful for businesses sending money across borders,
protects against stalkers,
protects against advertisers profiling you,
reduces identity theft,
prevents databreaches of personal info,
pushes forward cryptography,
allows people to purchase drugs (you decide if this is good or bad),
prevents financial censorship,
allows anonymous donations,
low fees,
more decentralized than bitcoin due to RandomX CPU mining,
prevents crypto robbery,
allows you to buy your adult content without anyone knowing.
large developer community iirc 3rd after bitcoin, eth
less volatile than other cryptos
usually most used crypto for payments when accepted at merchants
All of the pros - minus the inflation - also exist for Bitcoin-over-Lightning. Non of the cons exist for Lightning. Especially no waiting time. In lightning you can spend your received bitcoin after a split-second - no waiting whatsoever.
I had a terrible experience trying out Lightning. Could you tell me where I went wrong? I hit high fees converting to Lightning, and I hit high fees sending Lightning transactions.
The problem with Lightning or Zcash is that it is not private by default. A private currency has to be private by default, with the option to share info on the transaction—that's Monero.
Lightning is centralized, Chainalysis supports Lighning tracing—it does not support Monero tracing.
You could get the same advantages of Lightning on Bitcoin by making a layer-2 solution on Monero. Monero beats lightning at layer one.
Fees are not part of the technological stack, but the economics behind the providers you chose. Just as with the Internet, you can host your own webserver or datacenter at home and run your own AS in your bedroom. But most people chose to pay someone to do that. Right now regulation kills all sorts of crypto (Monero included) and drives up prices because it is mostly startups that provide services around that, and they are crushed by the paperwork.
pro: if the world devolves into an authoritarean hellscape, people will still have a form of undetectable currency until they find a way to get rid of this
con: this improves the chances if the world to devolving into an authoritarean hellscape
The taxpayer funding is often the smaller part the complete lifetime pay package.
> A 2014 study of U.S. Department of Defense appointees showed that 28% exited to industry. As of 2023, 80 per cent of U.S. four-star retirees are employed in defense industry.[0]
There are actually entirely reasonable, rational explanations for this, but it's not a great look.
Undoubtedly so! But blame the people who get free money out of your income to be impartial and make decisions, not the people who have to earn their pay to carry out the decisions. If they wanted to prohibit that sort of thing they could.
In my experience sonnet<opus by a long shot for code review. Sonnet often flags things as errors that are not, because it fails to grasp the big picture… and also fails to grasp structural issues that are perfectly coded and only show up as problems at the meta scale.
I have no reason to believe that the next generation won’t offer similar gains in verification, and there is some evidence to support that the cybersecurity implications are the result of exactly this expansion of ability.
It depends on how you review. In an orchestrated per-task review workflow with clearly defined acceptance criteria and implementation requirements, using anything other than Sonnet (handed those criteria and requirements) hasn’t really led to much improvement, but it drives up usage and takes longer. I even tried Haiku, but, yeah, Haiku is just not viable for review, even tightly scoped, lol.
Siccing Sonnet on a codebase or PR without guidance does indeed lead to worse results than using Opus, though.
That makes sense, if your scope is tight enough, good enough is good enough. I’ve got the expected specifications and code style guides, including some aerospace engineering ones, but in complex systems I still run into difficult to sus out corner cases where the code works but the system breaks, usually due to unresolved conflicts in operational requirements.
Lol yeah, I don’t think I’m ready to ride in the jet that Claude built lol. I should clarify that I use the code guidelines because they are solid guardrails for making things that perform predictably, not because I’m building MCAS lol. Let’s hope that “vibe aerospace engineering” is a way off for now.
That's what I meant to get at by "it runs on their cloud."
They can name that user-facing ultrareview API endpoint whatever they want, and we have no way to see what model endpoint it calls internally once running on their cloud, right?
Well, when the moltbook story was everywhere, later people thought it was some big gotcha that "oh, they were actually humans."
So, showing true agent to agent interactions is interesting, but one could never be sure that's what you were actually seeing unless you were in control of all the agents.
> Claude: "I don't have memory of previous conversations."
> You: "We spent 3 hours on it yesterday"
> Claude: "I'd be happy to help debug from scratch!"
People are twisting themselves into knots trying to solve issues like this. Let's use the mental model of "coding agents are not magic, treat them like humans." The boring old Jira MCP [0] completely solves the posed problem.
> You: "Remember that auth bug we fixed?"
> Claude: Called Atlassian Rovo
> Claude: "Yes, I see PROJ-123 with commit SHAs, comments on what we decided and why. How would you like to proceed?"
If all LLMs disappear, you still have human readable Jira tickets to continue working on your project, instead of something like QR codes in an MP4.
[0] I am not trying to promote Jira or MCPs here, use whatever you want. I went back to Jira because its usage patterns are very well represented in the training data, and their MCP is not in beta.
Thanks for suggesting this. Prior to going back to Jira, I had been curating my own set of .md docs & tickets in my repo, but it got way our of hand trying to roll my own.
Given Atlassian's AI training news, I will be trying out tk.
Yes, I go even further. In-repo, I have a chats folder that my /done skill fills with ~"what we did, and didn't accomplish in this chat. Blah blah (a few more instructions) - finish with a great hand off to the next chat to continue the work." I run that anytime I approach 50% of the context window, as all models get dumb at that point. Then /clear, then /effort max just to be safe, then "please ingest chats/2026-01-01-00-00-what-we-did.md and proceed." It's a very purposeful custom /compress that works far better in my experience. If I ever hit auto-compress, I have failed as a Claude jockey.
My friend, misanthropy is giving last year energy. Don't get so down. In other under-reported news: until the Earth reaches 100% de-population, it can in-fact get worse. Therefore, it's currently not that bad!
Let's take a moment to appreciate that we live on a populated planetary body. The Peter Thiel has not yet achieved its ultimate goal. Good times.
The amount of new revenue that I am personally able to create for my clients, using Claude models for dev, and Claude models inside the insanely agile products delivered, is astounding.
If I was not currently experiencing this myself, and someone told me that this was possible, I would be calling them names.
reply