Replying toNot stepping on bugs

Gauraventh6mo

I am glad you liked it! I find knowing when to post to LW very hard; judging from the downvotes I interpret that this post is probably not super relevant to LW users, but nonetheless happy you liked it.

Qualified rights for AI agents

Gauraventh

6mo

I expect to be wrong in some ways on this post. Would be keen for feedback!

Introduction/Summary

As AI agents increase in capability, conduct tasks over longer time horizons, and interact more deeply with the world, it will be worth asking whether such agents should be given rights? A primary argument against this is that AI rights may conflict with or undermine human rights. Once granted legal standing, AI agents would gain the protection of institutions, and the law could become a mechanism that favours AI interests over human ones. Highly capable systems might leverage legal protections, precedents, and enforcement in ways that constrain human autonomy, directly increasing the chance of loss of control... (read 1288 more words →)

Not stepping on bugs

Gauraventh

6mo

I let a tooth rot for over a year. It was my upper right second molar, the one just before the wisdom tooth. Every time I ate something hard, like a peanut, it would wedge into the decay and shoot a spike of pain through my upper jaw. I was lazy, so I ignored it until it got worse.

And then it did.

One afternoon in June, a truly terrible toothache set in. Before I could get an emergency dental appointment, I found myself standing in the shower, letting hot water run over my head to dull the pain. Afterwards, I planted my face into a cuddle bear on the sofa. I’d been told... (read 598 more words →)

My kids won't be workers

Gauraventh

6mo

Illustration by Pallavi Yadav

I previously wrote a post called ‘Wacky’ and deleted it. This is another attempt to explain why I think the future could get wacky, written mainly from the perspective of having children and the world they might grow up in. This is speculation; my biases and circles shape these views, and I may be wrong.

If I have children, it will likely be after artificial general intelligence (AGI) arrives. I would want to see its effects first. On current forecasts, that puts the decision in my 30s. By then I expect to be semi-retired, with most labour automated.

Some AI firms openly aim to automate all work, and I suspect they... (read 1583 more words →)

Gauraventh7moQuick Take

From Meta: https://www.meta.com/superintelligence/

Personal Superintelligence

Over the last few months we have begun to see glimpses of our AI systems improving themselves. The improvement is slow for now, but undeniable. Developing superintelligence is now in sight.
It seems clear that in the coming years, AI will improve all our existing systems and enable the creation and discovery of new things that aren't imaginable today. But it is an open question what we will direct superintelligence towards.
In some ways this will be a new era for humanity, but in others it's just a continuation of historical trends. As recently as 200 years ago, 90% of people were farmers growing food to survive. Advances in technology have

... (read 498 more words →)

Replying toLegal Supervision of Frontier AI Labs is the answer.

Gauraventh9mo

Legal Supervision of Frontier AI Labs is the answer.

Hey, thanks for your comment. I do think this is right. The style of writing on my blog is a lot more bullish, and does not index uncertainty. I appreciate this is not in the ethos of LW, so I am going to change how/what I crosspost here. I prefer being bullish because this seems to get a lot more feedback and good questions such as yours.

As for your questions - I intend to write a second part and address these. It is much easier for me to say something is needed versus actually trying to elucidate on what this looks like. So I expect to go do more work on this and return with a more fleshed out theory. In the meantime you may enjoy reading Peter's paper on supervision: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5122871

Legal Supervision of Frontier AI Labs is the answer.

Gauraventh

9mo

If the biggest threat model from AI systems comes from internal deployment, then the correct governance move is to establish independent legal supervisors for frontier AI labs^[1].

Steven Adler recently argued against relying on a "race to the top," where frontier labs compete to be the safest when deploying models.

‘A race to the top can improve AI safety, but it doesn’t solve the ‘adoption problem’: getting all relevant developers to adopt safe enough practices.’

We should be sceptical of actors racing to build what could become the most powerful technology in human history and then saying they'll compete to be safe. There’s already plenty of evidence for this. Google didn’t release model evaluations for... (read 861 more words →)

Gauraventh's Shortform

Gauraventh

10mo

This is a special post for quick takes (aka "shortform"). Only the owner can create top-level comments.

Replying toMATS mentor selection

Gauraventh1y

MATS mentor selection

Wow, my intuition was that it was really hard to get mentors onboard with supervising scholars given time constraints most senior researchers have, so seeing 87 mentors apply feels is wild!

Replying toCalifornians, tell your reps to vote yes on SB 1047!

Gauraventh1y

Californians, tell your reps to vote yes on SB 1047!

The Bill has passed the appropriations committee and will now move onto the Assembly floor. There were some changes made to the Bill. From the press release:

Removing perjury – Replace criminal penalties for perjury with civil penalties. There are now no criminal penalties in the bill. Opponents had misrepresented this provision, and a civil penalty serves well as a deterrent against lying to the government.
Eliminating the FMD – Remove the proposed new state regulatory body (formerly the Frontier Model Division, or FMD). SB 1047’s enforcement was always done through the AG’s office, and this amendment streamlines the regulatory structure without significantly impacting the ability to hold bad actors accountable. Some of the

... (read more)

CultFrisbee

Gauraventh

Tom and I have been pondering how to make friends as adults. One solution he has invented is CultFrisbee™. Here's how you play:

Rules:

You need at least one other player.
Before catching a frisbee, you must clap.

Fouls:

Not catching a 'properly' thrown frisbee
Failing to make a 'proper' throw
Not clapping before catching the frisbee
Clapping when a frisbee isn't thrown to you

Every time you foul, you get a letter, which spells out the word "Cult." Your first foul earns you a C, second a U, and so on. If you complete the word "Cult," you must approach someone and ask them to join a game of CultFrisbee.

Today, we managed to meet about 15 different people at the Glasgow Botanical Gardens and found two who gave us their numbers to stay in touch.

Replying to(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Gauraventh2y

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Fixed!

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Gauraventh

This was written by Gaurav Yadav and Robert Reason for the Bristol AI Safety Centre.

Thank you to David Staley, Gabriel Kulp, Lucy Farnik, Oliver Guest, and Oliver Marks for their feedback and help on this report. The views here do not necessarily reflect the views of these individuals.

Abstract

In the following report, we examine the strategies Taiwan may employ to protect both itself and its semiconductor industry. We discuss four strategies: 1) Scorch the Fabs - Taiwan threatens or executes the destruction of their fabs to deter invasion or deny their use for China - we conclude that the credibility of this approach is limited and intractable; 2) Sabotage the Fabs - leveraging vulnerabilities in... (read 6370 more words →)

LESSWRONG
LW

LESSWRONG
LW

Gauraventh

CultFrisbee

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Legal Supervision of Frontier AI Labs is the answer.

Qualified rights for AI agents

Gauraventh

Qualified rights for AI agents

Not stepping on bugs

My kids won't be workers

Legal Supervision of Frontier AI Labs is the answer.

Gauraventh's Shortform

CultFrisbee

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Gauraventh

CultFrisbee

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Legal Supervision of Frontier AI Labs is the answer.

Qualified rights for AI agents

Gauraventh

Qualified rights for AI agents

Not stepping on bugs

My kids won't be workers

Legal Supervision of Frontier AI Labs is the answer.

Gauraventh's Shortform

CultFrisbee

(Report) Evaluating Taiwan's Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

Introduction/Summary

Rules:

Fouls:

Abstract