Yeah, you have to have some level of involvement. It's crazy that we are actually talking of 100% dev involvement to literally 0% involvement.
It seems to me that chances of misalignment are a function of how much time you use them. Eventually, it will happen.
> Yes I still vibe unit test and integration test, False sense of security. What's stopping the agents from doing an elaborate form of "Assert.IsTrue(true)" in some/most/all of the tests since you are vibing it and not…
Look at the COBOL devs. They are been called in their retirement years. You don't want that to be you. Life is short
Because safety isn't a priority at OpenAI and they had (have?) been falling behind Anthropic in the LLM race? Big fans of the saying: move fast and...
Arguably, if the website does not have a positive value and only negative, the logical thing is to shut it down and keep the Google Business profile imo. I assume she has been paying for the domain name all this years…
> It takes four years of education before becoming a taxi driver in London That doesn't apply to private hire vehicle drivers afaik.
> Food should be prepared with love You forgot about the sentient non-veg elements that also get added to it. Is that done with love too?
> I can't tell the difference so what does it matter? Eternal September. Just like the usenet groups
> it's advancing fast Actually, it (LLM models) has been advancing slower for a while. This is pretty much in the mainstream. And the current approach is facing all sorts of obstacles that are not exclusively…
> This is a very VC focused lens Just like mostly everything he writes about.
> But yes, there is an interesting change in the past decade, where everything new must be over-hyped. Yep, the good old "Our tech is making the world a better place" has gotten old and a new approach was needed. If you…
Difference is you are not a trillion dollar plus technology hyped as a harbinger of civilisational change.
> I'm of the opinion that we still have some discoveries on the mathematical side of the fence to go that will improve models further. That's assuming the infrastructure needed to develop models stays available…
Not exciting for anyone directly or indirectly invested in a frontier lab or its partners. And that is a lot of people, including you.
Isn't video generation expensive? Are you letting/making users generate series of videos without any plan? I think content depth and breadth would be good to check pre-generation at a minimum.
And that's fine. Not everyone has someone to vouch for.
> I don't want to be constantly worried about my own reputation. That is why you must think carefully who you vouch for. Like in a house party. If you bring someone who does not act well, it reflects bad on you.…
Basically, any copyrighted material should do. This only covers not ablated LLMs though. Ultimately, a web of trust where someone whose humanity and integrity you trust AT THE COST OF YOUR OWN INTEGRITY vouches for an…
> Having no principles itself eventually loses you a small minority of users... and those were your core supporters and evangelists If pleasing that minority doesn't help Firefox gain more users (looking at the numers,…
> Simple: allow those users to be lost to preserve the product. Lose the users lose the funding. Firefox can't live out of HN as its sole userbase. Have you checked how much it costs to build and maintain a modern…
> We're choosing to call LLMs (and the little "harness" programs that query them in loops and execute their output) "AI", even though it doesn't make much sense. Most people are laypeople who have no idea what's on the…
> When you used google 10x a day, you saw, at a minimum, 20 ads/day That's assuming a user with no ad-blocker which is an ever diminishing number of users
> Does anyone want to use a tool that is capable of disguising an ad as a geniune recommendation at anytime? It depends on the economics and usefulness of the tool. Google Search got away with it because it was useful…
> (I want to spend no more than $10k. And I want to run a model comparable to today’s SOTA.) The question is, will you want to run a model comparable to today's (meaning 2026) SOTA in 2028? Humans always want the latest…
Yeah, you have to have some level of involvement. It's crazy that we are actually talking of 100% dev involvement to literally 0% involvement.
It seems to me that chances of misalignment are a function of how much time you use them. Eventually, it will happen.
> Yes I still vibe unit test and integration test, False sense of security. What's stopping the agents from doing an elaborate form of "Assert.IsTrue(true)" in some/most/all of the tests since you are vibing it and not…
Look at the COBOL devs. They are been called in their retirement years. You don't want that to be you. Life is short
Because safety isn't a priority at OpenAI and they had (have?) been falling behind Anthropic in the LLM race? Big fans of the saying: move fast and...
Arguably, if the website does not have a positive value and only negative, the logical thing is to shut it down and keep the Google Business profile imo. I assume she has been paying for the domain name all this years…
> It takes four years of education before becoming a taxi driver in London That doesn't apply to private hire vehicle drivers afaik.
> Food should be prepared with love You forgot about the sentient non-veg elements that also get added to it. Is that done with love too?
> I can't tell the difference so what does it matter? Eternal September. Just like the usenet groups
> it's advancing fast Actually, it (LLM models) has been advancing slower for a while. This is pretty much in the mainstream. And the current approach is facing all sorts of obstacles that are not exclusively…
> This is a very VC focused lens Just like mostly everything he writes about.
> But yes, there is an interesting change in the past decade, where everything new must be over-hyped. Yep, the good old "Our tech is making the world a better place" has gotten old and a new approach was needed. If you…
Difference is you are not a trillion dollar plus technology hyped as a harbinger of civilisational change.
> I'm of the opinion that we still have some discoveries on the mathematical side of the fence to go that will improve models further. That's assuming the infrastructure needed to develop models stays available…
Not exciting for anyone directly or indirectly invested in a frontier lab or its partners. And that is a lot of people, including you.
Isn't video generation expensive? Are you letting/making users generate series of videos without any plan? I think content depth and breadth would be good to check pre-generation at a minimum.
And that's fine. Not everyone has someone to vouch for.
> I don't want to be constantly worried about my own reputation. That is why you must think carefully who you vouch for. Like in a house party. If you bring someone who does not act well, it reflects bad on you.…
Basically, any copyrighted material should do. This only covers not ablated LLMs though. Ultimately, a web of trust where someone whose humanity and integrity you trust AT THE COST OF YOUR OWN INTEGRITY vouches for an…
> Having no principles itself eventually loses you a small minority of users... and those were your core supporters and evangelists If pleasing that minority doesn't help Firefox gain more users (looking at the numers,…
> Simple: allow those users to be lost to preserve the product. Lose the users lose the funding. Firefox can't live out of HN as its sole userbase. Have you checked how much it costs to build and maintain a modern…
> We're choosing to call LLMs (and the little "harness" programs that query them in loops and execute their output) "AI", even though it doesn't make much sense. Most people are laypeople who have no idea what's on the…
> When you used google 10x a day, you saw, at a minimum, 20 ads/day That's assuming a user with no ad-blocker which is an ever diminishing number of users
> Does anyone want to use a tool that is capable of disguising an ad as a geniune recommendation at anytime? It depends on the economics and usefulness of the tool. Google Search got away with it because it was useful…
> (I want to spend no more than $10k. And I want to run a model comparable to today’s SOTA.) The question is, will you want to run a model comparable to today's (meaning 2026) SOTA in 2028? Humans always want the latest…