Chris Lonsberry

Senior Applications Consultant & IT Generalist @ DLL Solutions
34 karmaJoined Working (6-15 years)Cleveland, OH, USA
aisafety.quest

Bio

Participation
2

  • Launching in 2023 > EA Cleveland - stay tuned.
  • Currently organizing > AI Safety Quest - We help people navigate the AI Safety landscape, build the community, and shorten the time to upskill and contribute.

How others can help me

  • Mentor me!
    • Help me be a better coach and advisor. 
    • Cybersecurity (as an IT Generalist rather than SWE or product builder). Industrial Control System experience would be a huge plus.
  • Improve my perspective on the AI Safety landscape. 
  • Best practices for managing and optimizing and org that relies on volunteers for their core mission.

How I can help others

  • Plan and achieve Financial Independence (FIRE), or whatever financial runway you need to be maximally effective.
  • I have experience working with a wide variety of process industries (chemicals, oil and gas, bulk transportation, energy) when I worked with OSIsoft PI System (owned by AVEVA)

Comments
13

Topic contributions
1

To pile on: a local Mexican chain also consistently puts less of PMAs into their tacos than they do meat. 

I'm generally a "hit the 1.6g/kg/day target" kind of guy and have been ignoring BCAAs. The effect of branched-chain amino acids supplementation in physical exercise: A systematic review of human randomized controlled trials - ScienceDirect suggests that's still a valid strategy, but I'm open to learning more.

Truly shocked by the research above. Another reminder of just how far one drifts from the norm when getting enthusiastic about a particular topic. I really thought (despite personal experience to the contrary) that everyone knows cottage cheese is high protein and peanut butter has some protein, but not great macros (high fat:protein ratio). 

I think we've drifted well away from the habits of normal consumers now, but I will add say that there seems to be some agreement that protein intake up to 1.6 grams per kilogram of body weight per day (g/kg/day) tends to maximize resistance exercise-induced gains in muscle mass. 

There's probably some room to lower that by looking for a point of decreasing marginal returns, but it seems to me that most weight lifters will target 1.6 + some safety margin that makes them feel good. 

My main point here is that among weight lifters who are paying attention to science, there's a very clear answer to how much protein we need (or choose to consume for gains).

I'll share my view on this as a consumer. Almost every time I see a PMA on a menu, I expect to get the following things if I order that option:

- Smaller quantity
- Worse macros (That is, I expect that the PMA will have a higher fat/carb:protein ratio than chicken breast)
- Often, an upcharge to boot

This expectation might be irrational on my part; an aftereffect of being vegetarian in the 1990s before the current generation of PMAs was released. When I was a full-time vegetarian, I frequently observed restaurant meals for vegetarians contained both less calories and less (sometimes almost no) protein. I guess the upshot of the above is that I believe achieving price- and nutrition- parity would strongly influence me.

Finally, I think I tend to do better at making choices that align with my ethics at the supermarket rather than at restaurants. 

I removed the Blue Dot AISF link.

> We no longer maintain an AI safety job board.

Copying my comment from the EA Forum version of this thread

Mentorship is one of the most frequently requested services that AI Safety Quest sees when conducting Navigation calls. I hope this service can help bridge the gap between "I want to do something about AI safety" and "I'm working on a meaningful AIS project". Many thanks to you both for making this happen.

Does anyone have a link to the slides? I could use more zoom 😅

This was posted to lesswrong 4 days ago; consider checking out the comments to the lesswrong version. There are 18 comments at the time I'm writing this.

"but why hasn't EA done this already?"

still seems like a fair question. I think the underlying problem you're pointing to might be that people will then give up on their projects or ideas without having come up with a good answer. An "EMH-style" mindset seems to point to an analytical shortcut: if it hasn't already been done, it probably isn't worth doing. Which, I agree is wrong. 

I still think EMH has no relevance in this context and that should be the main argument against applying it to EA projects. 

I had to use Wikipedia to get a concise definition of EMH, rather than rely on my memory:

The efficient-market hypothesis (EMH) is a hypothesis in financial economics that states that asset prices reflect all available information. A direct implication is that it is impossible to "beat the market" consistently on a risk-adjusted basis since market prices should only react to new information. [1]

This appears to me to apply exclusively to financial (securities) markets and I think we would be taking (too) far out of its original context in trying to use it to answer questions about whether great EA projects exist. In that sense, I completely agree with you that:

it's a poor way to model the situation that will lead you to make systematically wrong judgments

 In the real (non-financial) world, there are plenty of opportunities to make money, which is one reason entrepreneurs exist and are valuable. Are you aware of people using EMH to suggest we should not expect to find good philanthropic opportunities?

  1. ^
Load more