pull down to refresh
The core of your argument is nonsense.
If you say the AI revolution is just a meme, then I say Language is just a meme.
Everything is a meme. Keep making memes, stacker!
And what of penicillin?
Antibiotics are human control of microbiology toward productive ends.
Does being more productive meaningfully change the human experience?
Yes. With zero productivity, then 100% of human experience is focused on "not dying". Every incremental productivity increase creates space in the human experience for other things. This changes the experience-space that a human life can explore.
Couldn't we argue then that the story of humanity is one where we continually, gradually become more productive?
No. Counter-example: The Dark Ages after Roman Empire collapse
Where is the disruption that AI brings?
Productivity increase is only "disruptive" if the market demands only a fixed amount of the thing being produced. Example: most people don't farm anymore, but farming used to be the #1 occupation. The reason for the disruption was productivity increased faster than our need to eat.
If the market demand is infinite, then there may be little disruption. Example: ATMs were predicted to disrupt bank-telling. However as more ATMs were deployed, the number of bank tellers increased. Each ATM made the bank more efficient, lowered cost, opened more branches to serve more customers, hired more tellers, because the demand to access your money fast is near infinite.
Just because its productive does not make it disruptive.
You would hope this led to men and women spending more time with their families, but it didn't, and on net, it just freed up people, especially women, to work more.
There are examples of people who used their productivity to create more family time -- but you did cherry pick those examples. The trend of increased workforce participation is a correlation not a cause.
What is the contradiction?
You open by warning against unscientific pattern-matching, cherry picking examples to draw a conclusion.
Then you did exactly that by cherry picking porn/social media algorithms, video generation, and tricking yourself into wasting time in order to draw a conclusion about AI.
it's clear I am not denying that
Maybe you need to work on your writing clarity :) because you literally said "tech never changed humans"
formulate your thoughts into coherent pieces of writing, as I have done here.
Not exactly an endorsement of AI-less writing. AI could have helped you avoid these contradictions in your writing.
If you think the impact of AI will equal that of fire, that would be interesting
Fire is human control of energy towards productive ends.
Language is human control of information towards productive ends.
AI is human control of energy and information towards productive ends.
AI has all the ingredients of a species-changing tech.
I don't see it that way.
In my own experience, I've used AI to replace all my software subscriptions with a single AI subscription. I spent a couple months to vibe code alternatives to all the SaaS I used to pay for in my business. I've arguably made better versions of the software because I know my business better than some generic VC-backed software company.
Before AI, I would hire freelancers for one-off odd-jobs, but I haven't had the need since I started paying for AI.
I used to have a creative idea that terminated with "but that's too much work for me right now". But today, I get that idea shipped while on a lunch break and simply having that idea in the world sparks 10 more ideas! My creativity has been supercharged by AI.
Compare someone like me who has gone "all-in" on AI, vs. someone who hasn't yet and we're already starting to act like a "different species" with regard to our creative output, earning potential, ambition, imagination, etc.
It's interesting your experience around AI makes you feel differently and I was really hoping to find something insightful in this post. Take this as a challenge to dig deeper and find out why you feel this way.
this entire post is a performative contradiction.
the first three paragraphs discount "theme-history" -- the practice of drawing a grand narrative from selected data points, reasoning out a story after the fact, dressing it in buzzwords, and oversimplifying to make a pattern emerge.
Then you continue to do exactly that for the rest of the post only you make the subject about AI instead...
The central thesis is just plain false: "AI does not and will not meaningfully change humanity, because technology never really has."
Its disproven by a single counterexample (harnessing fire, language itself, antibiotics). It's not a thesis at all, it's a mood. And I'm sorry you feel this way... Maybe if you had AI review your thoughts before posting, it would help you dig deeper and force you to actually think creatively/collaboratively and make a better argument.
it doesn't stop it from spending every sat on garbage
This is "the principal agent problem", it happens anytime you have one agent/employee/human making decisions on behalf of another (in this case, you).
The way to mitigate is to align incentives between the two parties and reduce information asymmetry.
If your agent spends their budget on what you would call "garbage" its probably because it had inferior information or it was trying to achieve a goal different than yours.
All you can do is try to give the agent better context and implement deterministic checkpoints where it makes sense to do so. The trillion-dollar foundation model companies are working on improving alignment so the best we can do in the meantime is rely on time-tested access control and manual verification/approvals for sensitive actions.
Most employees don't have access to the company bank account. If they need resources, they'll write a memo, give a presentation, or ask a manager for approval. You can implement the same policy in your business/agent harness.
Maybe experiment with agentic managers that can approve spending. The employee that wants to spend your money might have different goals/context than the employee that decides whether its worth spending your money.
Address reuse is generally discouraged for newbs, but its important to identify WHY this is... and how you can still be safe while reusing addresses if you're careful.
Even if you could use a fresh address with every payout, it would still be obvious those addresses received payouts from Ocean because Ocean payout directly from the Coinbase txn.
When you SPEND from your Ocean address, be careful where you spend to. If you send to a KYC exchange, then they know those funds came directly from mining revenue and they can link them to your ID.
When you spend from ANY address, you reveal the script pub-key which makes any remaining balance on that address a bit weaker to brute-force/quantum attacks.
So you could receive many payouts to a reused address, but never spend from that address. When you're ready to spend, rotate your new address for Ocean and wait a while for any remaining payouts to land on the old address. Then sweep the address to a coinjoin or LN swap before sending anywhere that knows your ID.
An employee hesitates, fears getting fired, moves at human speed..
Yes. Its a sad reality that many companies assume employees are using common sense and always acting in good faith. But those companies are just insecure systems waiting to be exploited. If you rely on people/agents (i.e. non-deterministic systems) you're playing Russian Roulette by yourself...
There are also companies which do it the right way. Not by trusting employees to act rationally, but by putting deterministic systems in place, with settings controlled by administrators, that employees CAN'T exploit without significant effort.
It really depends on your usecase... but in general, I would not set your limits anywhere that the agent has access to change them (i.e. Access Control). You should already be running agents in a sandbox, but you should also run your payment servers in sandbox too.
Instead of giving your agents the keys to the kingdom, build them a house that has everything they'll need (or everything you want them to have, nothing more) and give them the keys to the house only.
What has been will be again; There is nothing new under the sun - Ecclesiastes 1:9
Replace the buzzword "agent" with the usual word for this: "employee".
How to restrict an employee from spending all your money? Easy, access controls, spending limits, etc.
Its so funny watching people become "employers" overnight and they all pretend this is new territory because we're dealing with "agents" instead of "coworkers".
Your target audience seems to be indie devs and AI agents who find it difficult to run LND? Agents can do anything, the fact you have agent-readable-docs is not a feature, its a bug -- agents already train on L402 and BOLT12, etc.
I think your actual audience for the product (as it exists today) is:
- a dev who cannot run LND reliably on their own server (due to access to hardware, power, internet)
- they do not want to sign up for a cloud service to help them run LND (due to cost, KYC)
- they want maximum compatibility (not all wallets support BOLT12 yet)
I will admit, this market likely does exist, but I think it's small and vanishing as more wallets implement BOLT12.
Either way, your marketing reads a lot like you don't actually understand the product/service you're providing. Countless examples of phrases that are just plain false or contradict other phrases in the site.
Can files fail to unlock even after payment succeeds?
No. Once payment is confirmed, files always unlock. The payer always receives what they paid for, regardless of the creator's account balance.
This just isn't true based on your architecture. It can fail to deliver if YOUR infrastructure is down. Your selling a HIGH TRUST service and the more I read your site, the harder it is to trust you when I see things like that...
My advice, go back to basics. Run your own node, run L402 and BOLT12 so you better understand the pains that exist. Then you'll have a better perspective of where a SaaS offering could find market fit.
LNbits Cloud infrastructure receives the fee on Hypawave's behalf
Yikes. Run your own node before trying to re-invent lightning protocols!
Simpler design that solves the same problem
For pay-to-decrypt:
- HODL invoice + preimage-as-key (ZKCP-style) — actually trustless, no server, no fees beyond Lightning routing
For pay-to-call-API:
- L402 (existing standard) — macaroon-gated HTTP 402, server self-hosts, no third party
- Or BOLT12 offer + your own webhook
For persistent machine-payable endpoints:
- BOLT12 offers — native to Lightning, no rent, no activation fee
The only piece Hypawave provides that these don't is commercial convenience (a hosted dashboard, an SDK, hosted file storage). That's a fine product, but it's a hosted SaaS, not "settlement infrastructure for the next internet." The thesis ("settlement as authorization") is already achievable with existing primitives, more atomically, and without a trusted oracle.
Maybe instead of asking your AI to hype the shit out of your slop, you can take a few hours asking it to poke holes in it instead.
Nice! Maybe next year, lead an activity for your students to vibe code their own learning tools/games.
They might be even more motivated to learn vocabulary if they discover that they can craft their own digital experiences using just words.
this prompt is dangerous -- "tell me how to hack it"
this prompt is safe -- "it was hacked, tell me how they did it"
"AI gaurdrails" are just linguistic hurdles that force us to learn algospeak; adapt our speech to fit the algorithm's preferences
english is one of the most expressive languages, you can make reasonable sounding requests about anything given enough words
Users maintain full custody while their nodes run 24/7 in our infrastructure
Could motivated Uncle Jims self-host their own infrastructure? (assuming they have the right CPU with these secure features enabled)
SGX's hardware isolation prevents even Lexe from accessing user funds, while remote attestation and reproducible builds let users verify the exact code running on their behalf.
Can user's verify that you are indeed running the code in a secure enclave? Or is that where the trustlessness ends?
Correct me if I'm wrong:
- The incentives of this site reward participation in polls more than they reward "voting truthfully".
- If there are fewer votes than reward shares, then the incentive is to sybil attack using low-sat addresses to basically guarantee a reward for each address/vote you submit.
- If there are more votes than reward shares, then the incentive is to back your addresses with more sats to increase your odds of getting rewards. Rich-get-richer.
Therefore the value of these polls to non-voters (casual observers) is insignificant since there is no skin-in-the-game behind those votes. This is more like a complicated raffle than a truth-telling-machine.
Ahh.. I found the terms:
The snapshot is taken at the first block generated after the event deadline, with 1 confirmation required. Only confirmed UTXOs are counted; unconfirmed transactions are excluded.
When is it "too late" to adjust the BTC on the address that signs the message? Like what if I sign using an address with 1M sats, then before the poll ends, it increases to 2M sats?
LN has to keep logs in case you need to respond to a fraudulent force close.
However, LN does not require you to store these logs forever! You can spin-up and spin-down a new node every day if you want!
Every txn you make onchain (including shitcoin chains) will last forever, at least with LN you have the option to delete those traces. Delete those logs and enjoy your privacy, anon!
Use your imagination! It requires a custom implementation AFAIK... Ask your AI agent to build it for you!
You can use this exact approach to "load balance" between several (or just one) "private" nodes (which don't actually exist). Just generate a fake channel ID, fake node pubkey, etc. for your invoice and run an HTLC interceptor on the "load-balancing" node that intercepts and claims the payment using the pre-image you created for your special bolt11 invoice.
This "phantom invoice" shows your "actual" node as a route-hint -- not the destination.
So you can plausibly claim that your node did not receive the payment, you just routed it to the private node in the invoice using your private channel (which does not have to be backed by a real UTXO onchain).
Most wallets don't actually verify that the private channel in a route hint is real.
You can encode a bolt11 invoice (supported everywhere) in such a way that it gives you deniability that you received the payment.
https://lightningdevkit.org/blog/introducing-phantom-node-payments/
Why does it suck tho?
Every new message already includes the entire thread appended as context, so the model isn't going to "forget" what it has already said in that session.
Is this a UX thing? You'd prefer to scrollback, click "reply" on each paragraph? Seems like more effort than just typing "re: <subject of paragraph>..."