Cursor and Claude are a lot better than Copilot, but none of them can be trusted. For existing large code repos, LLMs can generate tests and similar boring stuff. I suspect there’ll be an even bigger shift to micro services to make it easier for LLMs generate something that works.
It’s similar to fixing code written by interns. Why hire interns at all, eh?
Is it faster to generate then debug or write everything? Needs to be properly tested. At the very least many devs have the perception of being faster, and perception sells.
It actually makes writing web apps less tedious. The longest part of a dev job is pretending to work actually, but that’s no different from any office jerb.
Yes, you can have multiple devices with the same seed for the pseudorandom number generator. You can turn any computer into a hardware authenticator. In practice, it depends on the bank or your employer. Google reduced phishing success rate to zero after switching to ubikey.
As for perception, you really nailed it. It’s more important than actual difficulty of gaining access to your accounts. Remember that most articles are written by low skill blue teamers who manipulate your perception into thinking it’s really easy while they don’t possess the skills to do it. Always call them out in a manner like “you claim it’s easy, have you done it?”. They will always say no.
Grape, my nigga.
Start bulking up by eating well, solid exercise routine, a bit of help from anabolic steroids. Pose with a formula-filled blackboard background shirtless while flexing your biceps for Instagram and Twitter. Become the math bodybuilding icon. Make jokes like “my muscles are not differentially equal to yours”. You should build an audience, and after that you’ll be able to expand into sponsorships, and OnlyFans. You can also do IRL prostitution, and earn thousands of $ per night. The key is to target either old hags, or rich homosexuals.
Good luck. Let your biceps look like the bell curve of a Gaussian distribution
Large context window LLMs are able to do quite a bit more than filling the gaps and completion. They can edit multiple files.
Yet, they’re unreliable, as they hallucinate all the time. Debugging LLM-generated code is a new skill, and it’s up to you to decide to learn it or not. I see quite an even split among devs. I think it’s worth it, though once it took me two hours to find a very obscure bug in LLM-generated code.
With current kWh/token it’s 100x of a regular google search query. That’s where the environmental meme came from. Also, Nvidia plans to manufacture enough chips to require global electricity production to increase by 20-30%.
Proton gives data to governments if requested. Why are you trying to shill it?
Pedantic types always mention that secure is only relevant in the context of a particular threat model. The elderly can use hardware authentication like those RSA devices or ubikey. Unfortunately, this is expensive, and banks don’t believe there’s demand for that. Would you switch banks for this feature?
What about people who only have one device? Kids, elderly, people with only work computer.
Simply paying is not sufficient. You need to be a telecom company, or a researcher afaik.
In what world would the US gov care to get into your bank account? Or your Facebook account when it’s already tightly controlled?
Proton is already used for identity management: OTP via email. They’ll implement OAuth if there’s enough demand for it. A company’s purpose is to be profitable, ethics side is largely irrelevant.
Many countries already have digital government ID: Australia, Estonia, Russia.
Watch the video again to see how hard it was for Derrick to get access. He got it via his telecom/academia researcher contact.
S7 will be retired or extended with access control. TOTP apps don’t work for edge cases like broken phone. Dedicated token devices get lost. SMS will continue being the main solution for 2FA.
Large gains were due to scaling the hardware, and data. The training algorithms didn’t change much, transformers allowed for higher parallelization. There are no signs of the process becoming self-improving. Agentic performance is horrible as you can see with Claude (15% of tasks successful).
What happens in the brain is a big mystery, and thus it cannot be mimicked. Biological neural networks do not exist, because the synaptic cleft is an artifact. The living neurons are round, and the axons are the result of dehydration with ethanol or xylene.
Not true. SMS is encrypted in 3G, LTE, 5G. Block cyphers like Kasumi and A/9 are used. SMS is reasonably secure, because it’s hard to infiltrate telecom systems like S7
AGI or human level intelligence has a hardware problem. Fabs are not going to be autonomous within 20 years. Novel lithography and cleaning methods are difficult for large groups of humans. LLMs do not provide much assistance in semiconductor design. We are not even remotely close to manufacturing the infrastructure necessary to run human level intelligence software.
LLMs are not programmed in a traditional way. The actual code is quite small. It mostly runs backprop, filters the data. It is already easily generated by LLMs.
Broke back convolution
To make sense of that, figure out what pays more observing/editing or cooking/writing. Big shekels will make boring parts exciting