> If you give a robot the ability to delete production it’s going to delete production
If you give an intern the ability to delete production, it's going to delete production. But to be honest you can as well replace "intern" or "robot" by human in general. Deletion in production should have safety layers that anyone cannot accidentally do it, specially without the ability of rolling back.
That’s a broken analogy. An intern and a llm have completely different failure modes. An intern has some understanding of their limits and the llm just doesn’t. The thing, that looks remarkably human, will make mistakes in ways no human would. That’s where the danger lies: we see the human-like thing be better at things difficult for humans and assume them to be better across the board. That is not the case.
I think the difference, though maybe I'm incorrect, is that when we have interns on our codebase they get restricted permissions. Can't push to prod, need pull requests with approvals and reviews, etc. Certainly can't delete backups. Whoever setup the robot's permissions did it wrong. Which is interesting because early on there were people complaining that these AIs refused to push to main, but now this stuff keeps happening.
> Whoever setup the robot's permissions did it wrong.
It doesn't have permissions of it's own. The way he's using it, it has his permissions.
Also, in order to be able to do deployments like that you need pretty wide permissions. Deleting a database is one of them, if you're renaming things for example. That stuff should typically not happen in prd though
I had a senior tech lead delete production doing a late night quick fix. Especially in panic mode where sometimes processes are ignored, things are going to go wrong. Don't need interns for that, nor llms.
> Didn't Proton already say that they were physically relocating their servers outside of Switzerland because the Swiss government couldn't be trusted?
They said they want to relocate to Germany which I would say in a polite way, is much worse in this regard.
In what sense? Germany has among the strongest judicial oversight for invasion of privacy in Europe. Due process is followed when securing search warrants that provide access to subscriber data (Germany does not have administrative subpoenas like in the US and other countries).
Former attempts at surveillance have been struck down in the Bundesverfassungsgericht, and the right to privacy has even been affirmed for foreigners (as opposed to other countries like the US that reserve that foreign nationals have zero due process rights for invasion of privacy).
Germany is an absolutely terrible choice for this. Other Email providers such as Tuta which also offer encrypted emails, were forced to install a backdoor. As soon as the police arrive, every future email sent to the account in question is copied unencrypted without the person being informed.
This is much worse than passing on payment details or stored backup email addresses, as Proton Mail is required to do in Switzerland.
> Other Email providers such as Tuta which also offer encrypted emails, were forced to install a backdoor. As soon as the police arrive, every future email sent to the account in question is copied unencrypted without the person being informed.
Important caveat: Tuta was required by a court to provide police with access to a customer's _unencrypted_ emails (ie regular SMTP mail). The police had also asked for a backdoor to Tuta's E2E emails, and that request was rejected by the courts.
But the idea behind Tuta and Proton is that emails are encrypted when they arrive in the inbox. The fact that emails sent between Tuta users are still safe offer little added value because distribution is far too limited. The reason people choose such a provider is that they do not want the authorities to have access to their mailbox, but this is undermined by a backdoor. Switzerland is much better off in terms of the legal situation in this area.
In the sense that it's a joke that caves in to the flimsiest pressure from a certain superpower. Although pressure is a bad choice, it's more like it's a wholy owned subsidy.
I think the space of RSS feed readers and aggregators are very rich already. The pain point for ordinary users is to have easy way to generate RSS feeds for websites that don't provide organic one.
There are few options but mostly proprietary and expensive. And no normal person will want to play the CSS tricks to extract feed that something like FreshRSS support.
I disagree because yes I dislike the whole JS ecosystem and the language itself. But also because Electron apps in general are resource monsters and while some are better than the others, Claude Desktop is definitely not one of them. Hell even their website will crash on Firefox very often.
> OpenAI reportedly discussed charging $20k/month on PhD-level research agents with investors.
At this price point, it will be cheaper to hire a bunch of actual PhDs. The vast majority who will not earn anything close to 250k per year in most of the world.
I also seriously question what even does PhD-level mean in the context of a model? Someone with a PhD has developed a very deep but narrow knowledge of a particular domain and has contributed to at least pushing out our sphere of knowledge a tiny bit in that pillar of competency. A model is a best a brittle, fractured and often inconsistent representation of written human knowledge and lacks most basic intuitive grounding in the world due to the lack of embodiment.
In my experience, to safely get any value out of an LLM, you have to be more knowledgeable than the LLM on a topic. So in this case, you'd really need a PhD to use this tool, so at best its a $20k a month research aid, which honestly is far more expensive than a handful of grad students, and probably less effective.
Yes, but I suspect part of the pitch is that the “PhD level models” will accomplish 50x as much in a year as the human PhD, because they’re faster AND they work 24/7.
Whether they can deliver is another question, but I wouldn’t bet my career that they can’t.
its surprisingly easy to get away with murder (literally and figuratively) without piercing the corporate veil if you understand the rules of the game. Running decisions through a good law firm also “helps” a lot.
A bit over five years ago, someone struck and killed my friend in a crosswalk. He was a fellow PhD student. It was on a road with a 30mph limit but where people regularly speed to 50+mph.
He was an international student from Vietnam. His family woke up one day, got a phone call, and learned he was killed. I guess there was nobody to press charges.
She never faced any accountability for the 'accident'. She gets to live her life, and she now runs a puppetry education for children. Her name even seems to have been scrubbed from most of the articles about her killing my friend.
So, I think about this regularly.
I was a cyclist at the time so I was aware of how common this injustice was, but that was the first time it hit so close to home. I moved into a large city and every cyclist I've met here (every!) has been hit by a car, and the car driver effectively got only a slap on the wrist. It's just so common.
> Her name even seems to have been scrubbed from most of the articles about her killing my friend.
I'm somewhat surprised there were even articles? Are road fatalities uncommon enough in the US that everyone gets written up? Or was this a special enough one?
Not sure if this is true for every university, but when someone in the community dies, especially a student, there's usually at least an article about it.
Well the important concept missing there that makes everything sort of make sense is due diligence.
If your company screws up and it is found out that you didn't do your due diligence then the liability does pass through.
We just need to figure out a due diligence framework for running bots that makes sense. But right now that's hard to do because Agentic robots that didn't completely suck are just a few months old.
I would be less interested in scattering amplitude of all particle physics concepts as a test case because the scattering amplitudes because it is one of the concisest definition and its solution is straightforward (not easy of course). So once you have a good grasp of the QM and the scattering then it is a matter of applying your knowledge of math to solve the problem. Usually the real problem is to actually define your parameters from your model and define the tree level calculations. Then for LLM to solve these it is impressive but the researchers defined everything and came up with the workflow.
So I would read this (with more information available) with less emphasize on LLM discovering new result. The title is a little bit misleading but actually "derives" being the operative word here so it would be technically correct for people in the field.
reply