I've been using both on a Rust codebase and have found both work fairly well. Claude code is definitely more capable than Gemini. What difficulties have you had?
The biggest pain point I've had is that both tools will try to guess the API of a crate instead of referencing the documentation. I've tried adding an MCP for this but have had mixed results.
It might be that we have multiple features in our codebase and Gemini seems to struggle understanding that it needs to be aware of #[cfg(feature = "x")] and also that if it's trying to run things, it might need to specify the feature.
And yes, when they guess APIs, it's highly annoying.
Congrats on the launch! This is really cool - one of the applications of LLM I find most compelling. I've seen so many back office processes that have hundreds of steps, are incredibly error prone, and traditionally couldn't be automated due to API limitations. Solutions like Skyvern are going to supercharge businesses that have had historically low margins due to the number of humans required. (Not as a replacement for a human, but as a force multiplier)
The most fascinating part is how tough that work really is. Everyone we've talked to loathes the manual stuff, but until a better solution comes out, you have to allocate X% of your time to it
Only if you require text message based two factor. Password managers like 1Password allow you to store your OTP within them and share that + the password internally within your team
I've set up multiple times a phone-to-Slack proxy for this exact reason. In my case it was a VoIP number, but if that's blocked, Android has many SMS-to-webhook apps and even entry-level industrial LTE routers generally have this feature so you can use a real SIM card.
This is really great. Congrats on shipping it! You might find https://www.lingq.com helpful as a source of inspiration. I think it's a fairly similar concept.
LingQ's killer feature for me is that as you click on words (or phrases - which I find really helpful btw) to translate them, they are added to your vocabulary list. It will automatically create flashcards for you from this list for SRS. Plus when you're reading a new story, words that are in your vocab list are highlighted yellow and new words are highlighted blue.
thank you! I will have a look at LingQ, it does look interesting. Some other people also asked me to add the vocabulary list for Webbu too, so that should be coming soon :).
There is a large graveyard of startups that have tried to crack open this market. It's a quite hard concept to execute on.
Some companies have made headway in eyewear and shoes (https://www.amazon.com/b?node=23595320011). Clothing is much more difficult - a lot of variety in body shape to accommodate.
So, making a buck creating a huge problem, and then making another buck working on the solution to that problem as a side-hustle. Boy, I'm really proud to be working in Tech in 2023...
Can an AI physically go to the DMV and get a REAL-ID? If we reach the point where this seriously is a concern, then we will have other, much larger problems to deal with. WorldCoin feel like a scam. The crypto aspect and then paying people in Africa to gaze into metal orbs and get their irises scanned (with funding from A-Z and Vinod Khosla), ... it's not exactly a confidence builder.
I actually think his actual concern here is humans stealing identities and running bots in order to farm any sort of UBI mechanism that gets set up.
It's not about humans asserting their humanness over some evil AGI/ASI machine, it's about sybil resistance. Malicious humans pretending to be thousands or millions of other humans.
People will soon be able to use their iriscodes to reclaim credentials that they sold, and once people start doing that, the market for re-sold credentials will drop to approximately zero.
> People will soon be able to use their iriscodes to reclaim credentials that they sold, and once people start doing that, the market for re-sold credentials will drop to approximately zero.
Is the obvious related problem this solution brings forth supposed to go unsaid? (Criminals don't buy something they can steal.)
Imagine we live in a future where Worldcoin is ubiquitous, all the hardware/software is open source. Several orb operators exist in every city in the world, and operators have a non-trivial amount of WLD tokens staked to ensure good behavior. Misbehaving orbs have their keys and subkeys revoked at the first sign of illicit usage.
Maybe a criminal kidnaps 10 people, then steals an orb, and uses those 10 iriscodes to register 10 new credentials on the network. How much profit would you expect to come to the criminal before their credentials are revoked?
> a criminal kidnaps 10 people, then steals an orb, and uses those 10 iriscodes to register 10 new credentials on the network
This is a lot of work. Just steal the iriscodes. If they're used for re-issuance, they're being stored and transferred. If they are not directly used, there is a hash-like reduction that can be exploited.
For sake of argument, though, let's assume perfect security. Infallible security. All the way through. Congratulations, you've turned every pair of eyeballs into an oil spigot. When the Taliban or ISIS carves through a town, instead of beheading the leaders and taking their treasures, they take everyones' irises. Every authoritarian state would require scans of its citizens so payment could be routed through (read: stolen by) it, a requirement they would back up with violence.
So if we start with your premise, the attack I can see being a concern is being able to deny people access to Worldcoin and burning their accounts/tokens by maliciously attacking an orb. If keys can be revoked because of illicit usage, someone could generate a key at the orb they wish to attack and act maliciously. Sybil attack, or attempt fraudulent transactions, or whatever behavior triggers burning all the keys tied to an orb. Do this before an election (if worldcoin is used for voting) or before the UBI is dispersed for the month and you could disrupt a lot.
How does Worldcoin provably know that an attacker is not using a dead person’s keys?
That sounds like a automation problem. If you have an office dealing with UBI and unemployment in your neighborhood, the government can literally know their customer. The more people unemployed, the easier it will be to staff them.
These people are pretending they're fighting fraud, when instead they're trying to reinvent government responsibilities/campaigns as passive income sources for 3rd party rent-seekers. Uber for welfare.
That sounds neat. I like the ultra-local approach.
The important thing that I think your argument is missing is that everything here is additive. I'm sure the whole Worldcoin team would love it if governance organizations of all sizes, from tiny villages to billion+ population nation states would run UBI programs to take care of their people. The thing is, today, that's not really happening, so there are a number of people working to make it happen from different angles. Maybe all of them fail, maybe one succeeds, maybe a bunch succeed and any given citizen of the world might be entitled to payments from a handful of UBI sources.
Do you think there are better global scale UBI projects and ideas out there, or do you think we should just stop trying to solve this problem?
Sybil resistance is important in any sort of UBI system that preserves privacy, which may or may not be important. I like the idea of a privacy preserving UBI, and feel like it's definitely a lot more equitable than having it based on existing passports, or bank account numbers, or something like that.
This is valid. But even if we reduce scope to a privacy-preserving UBI, nothing Worldcoin solves the core hurdles thereto. It's solving a side problem tangentially related to those involved with a privacy-preserving UBI, but entirely germane to an a16z-style crypto pump-and-dump.
Adjunct: how is Worldcoin privacy preserving? At least with Bitcoin you can plausibly deny ownership of a wallet. In a world on Worldcoin, every traffic stop lets the cops inspect your bank account.
The biggest pain point I've had is that both tools will try to guess the API of a crate instead of referencing the documentation. I've tried adding an MCP for this but have had mixed results.
https://github.com/d6e/cratedocs-mcp