Josh Blake
@joshuablake.co.uk
Professionally: civil servant, science advice for emergencies, previously infectious disease statistics PhD
Interests: Effective Altruism, politics, nerds
Interests: Effective Altruism, politics, nerds
Do the safety concerns here also extend to the built in web search tools, if you've provided sensitive information in your prompt?
September 13, 2025 at 1:35 PM
Do the safety concerns here also extend to the built in web search tools, if you've provided sensitive information in your prompt?
This is only employment income right? Don't the very highest earners get most of their income from capital gains?
August 30, 2025 at 8:18 AM
This is only employment income right? Don't the very highest earners get most of their income from capital gains?
Supposedly fixed
simonwillison.net/2025/Apr/29/...
simonwillison.net/2025/Apr/29/...
April 29, 2025 at 7:23 PM
Supposedly fixed
simonwillison.net/2025/Apr/29/...
simonwillison.net/2025/Apr/29/...
Supposedly fixed
simonwillison.net/2025/Apr/29/...
simonwillison.net/2025/Apr/29/...
A comparison of ChatGPT/GPT-4o’s previous and current system prompts
GPT-4o's recent update caused it to be [way too sycophantic](https://simonwillison.net/2025/Apr/28/sam-altman/) and disingenuously praise anything the user said. OpenAI's [Aidan McLaughlin](https://tw...
simonwillison.net
April 29, 2025 at 7:23 PM
Supposedly fixed
simonwillison.net/2025/Apr/29/...
simonwillison.net/2025/Apr/29/...
First one makes searching easier IMO
April 9, 2025 at 12:13 PM
First one makes searching easier IMO
Defer to CEA. If they accept you they think it's worth having you there. If you're worried about wasting others money then you can buy the full cost ticket.
March 14, 2025 at 7:59 PM
Defer to CEA. If they accept you they think it's worth having you there. If you're worried about wasting others money then you can buy the full cost ticket.
(I think I might be too low context on what argument is going on here too usefully engage so feel free to ignore)
February 6, 2025 at 7:06 PM
(I think I might be too low context on what argument is going on here too usefully engage so feel free to ignore)
I generally don't support making bad arguments to counter bad arguments
February 6, 2025 at 7:03 PM
I generally don't support making bad arguments to counter bad arguments
Argument proves too much IMO, basically lets you disregard all thought experiments where you don't like the outcome. Which I don't think you or Scott would agree with. (It's not really an argument, it's more rhetoric anyway.)
February 6, 2025 at 7:00 PM
Argument proves too much IMO, basically lets you disregard all thought experiments where you don't like the outcome. Which I don't think you or Scott would agree with. (It's not really an argument, it's more rhetoric anyway.)
This is because of my high respect for Rethink Priorities (which Peter was co-CEO of until recently), his forecasting track record, and the quality of a handful of old blog posts I've read by him. Not the current post which is mostly "coming soon".
January 31, 2025 at 7:09 PM
This is because of my high respect for Rethink Priorities (which Peter was co-CEO of until recently), his forecasting track record, and the quality of a handful of old blog posts I've read by him. Not the current post which is mostly "coming soon".
Do you have a good initial prompt to get an interesting conversation rather than an almost lecture each time you ask something?
January 27, 2025 at 7:52 AM
Do you have a good initial prompt to get an interesting conversation rather than an almost lecture each time you ask something?
@helenbranswell.bsky.social. But honestly, there's very little signal and most information is noise (unless you're trying to work out detailed policy).
January 24, 2025 at 5:59 PM
@helenbranswell.bsky.social. But honestly, there's very little signal and most information is noise (unless you're trying to work out detailed policy).
What do power calculations provide that confidence intervals don't? In both cases, you're asking how big an effect is compatible with the observed data.
January 5, 2025 at 6:19 PM
What do power calculations provide that confidence intervals don't? In both cases, you're asking how big an effect is compatible with the observed data.
This search strategy kinda reminds me of how Prolog works. But you've now got a massive fact base (training data) and a fuzzier goal (Raters)
January 5, 2025 at 8:18 AM
This search strategy kinda reminds me of how Prolog works. But you've now got a massive fact base (training data) and a fuzzier goal (Raters)
Reposted by Josh Blake
Here's a question I have NEVER managed to get a straight answer to: if I paste an internal company strategy document into ChatGPT or similar to ask for a summary, what are the chances that a competitor in six months time can ask the model about my strategy and get it to leak internal secrets?
January 4, 2025 at 11:24 PM
Here's a question I have NEVER managed to get a straight answer to: if I paste an internal company strategy document into ChatGPT or similar to ask for a summary, what are the chances that a competitor in six months time can ask the model about my strategy and get it to leak internal secrets?
Then there's the same question about the right combination of mutations arising though right? Somewhat complicated by the mutations possibly happening in non-human hosts.
December 30, 2024 at 4:15 PM
Then there's the same question about the right combination of mutations arising though right? Somewhat complicated by the mutations possibly happening in non-human hosts.
Can we apply this to a question roughly "at what point have we seen so many dairy workers infected without (enough to be detected) onwards spread that we should conclude this is so unlikely to cause an epidemic that we should focus efforts elsewhere"?
December 30, 2024 at 4:09 PM
Can we apply this to a question roughly "at what point have we seen so many dairy workers infected without (enough to be detected) onwards spread that we should conclude this is so unlikely to cause an epidemic that we should focus efforts elsewhere"?
2. More human infections (/increasing rate) is more rolls of the dice BUT each infection reduces our posterior probability of any infection becoming a pandemic. How do these shake out in overall update of risk? Probably not an original thought - maybe there's work on this already.
2/2
2/2
December 30, 2024 at 8:46 AM
2. More human infections (/increasing rate) is more rolls of the dice BUT each infection reduces our posterior probability of any infection becoming a pandemic. How do these shake out in overall update of risk? Probably not an original thought - maybe there's work on this already.
2/2
2/2
I do sometimes wonder if we're looking in the wrong places.
1. This paper (screenshotted) suggesting that purely zoonotic (no h2h transmission) don't become epidemic, it's the ones with small outbreaks that do.
academic.oup.com/mbe/article/...
1/2
1. This paper (screenshotted) suggesting that purely zoonotic (no h2h transmission) don't become epidemic, it's the ones with small outbreaks that do.
academic.oup.com/mbe/article/...
1/2
December 30, 2024 at 8:46 AM
I do sometimes wonder if we're looking in the wrong places.
1. This paper (screenshotted) suggesting that purely zoonotic (no h2h transmission) don't become epidemic, it's the ones with small outbreaks that do.
academic.oup.com/mbe/article/...
1/2
1. This paper (screenshotted) suggesting that purely zoonotic (no h2h transmission) don't become epidemic, it's the ones with small outbreaks that do.
academic.oup.com/mbe/article/...
1/2
What's does Target have that Tesco Extra or massive Sainsbury's don't?
December 26, 2024 at 7:34 PM
What's does Target have that Tesco Extra or massive Sainsbury's don't?
Should have tagged the author @binarybits.bsky.social as a consistently insightful voice on AI, and probably the best writer on self-driving cars.
December 23, 2024 at 11:17 AM
Should have tagged the author @binarybits.bsky.social as a consistently insightful voice on AI, and probably the best writer on self-driving cars.