2,159 karma · joined August 4, 2011
They cite the example of churches that don't receive post, I guess if somebody had a TV there & used it to watch live broadcast TV then the building would need a TV license. I do wonder when the TV Licensing door knockers would show up trying to gain entry... during a church service seems a little inappropriate!
1: https://www.theguardian.com/technology/2006/jul/13/epublic.g...
How come the separate database vs the Postcode Address File? Sounds like you could write an interesting "Things programmers mistakenly believe about UK addresses" post very easily :-)
Many of the papers go over my head at times of course (despite my interest, I'm firmly a layman!)
More generally, I've had a lot of success with searching for meta-analysis papers on topics I'm interested in, since the authors have already done the hard work of researching the literature and presenting core takeaways (as well as providing terms for more searches).
1: https://bsky.app/profile/abeba.bsky.social or https://twitter.com/Abebab
2: https://bsky.app/profile/olivia.science or https://twitter.com/o_guest
What an absolutely wonderful insight put extremely succinctly, thank you!
I read about a paper[1] a while back, it was a rather unpleasant animal study in cats. Using two kittens, one was free to look around, but the other they immobilised in some way, and made it see what the other kitten saw as it looked around. They discovered the immobilised kitten's visual processing did not develop normally whereas the mobile kitten's did, suggesting that it's not just the sensory input, but it being feedback to some internal agency within the brain.
I suspect the same thing would be true for attempts to develop AGI by giving them an audio-visual copy of a human's environment growing up: that internal state driving (and getting feedback from) the action to investigate/interact with the world is key.
1: https://arxiv.org/abs/1604.03670 "Interactive Perception: Leveraging Action in Perception and Perception in Action"
Another commenter has noted about the unpleasant sounds and ominous design, but I'm curious about your experience with other patient orientations - it occurred to me that a lot of the discomfort comes from going in head-first and on your back, that way you're losing any sort of autonomy.
I assume there are scans that require people be on their front and looking out from inside the machine (or otherwise able to see out of the machine directly rather than via mirror), do you find that patients for those procedures don't require sedation as often?
Isn't this especially dangerous on a build worker? All your source code goes in and you (presumably) use the binaries that come out across the rest of your infrastructure. Compromising a build worker in a persistent fashion due to lack of SIP seems like it could do some serious[1] harm...
I read the linked PDF and did some searching but I didn't see in what instances networks use a 5G Logo for LTE (although I do recall hearing about something like that in the past)... could you give some more detail?
With that said, there are still a lot of state-operated cameras. Mostly to make money from traffic infraction fines, curiously that's one area where governments don't ever seem to cut police funding :-)
>> Don't forget to disable sftp in your sshd_config >Doing `strace -fp $(pgrep sshd)` then `sftp userfoo@localhost` in another terminal, I see: [...] >execve("/tmp/custom-shell.sh", ["custom-shell.sh", "-c", "/usr/lib/ssh/sftp-server"]
If your sshd_config had "Subsystem sftp internal-sftp" as a shipped default, then I don't think it would have gone through the user's shell. Not a major issue, but there are just so many little points to be certain are locked down
>I'd love to see a `ssh/sftp userfoo@localhost` call that does something other than call cmd_foo, cmd_bar, or erring out.
The reference to LD_PRELOAD was to hint at that (since if the attacker can get a binary on the system -- e.g. in a git repo or config file) they could change what functions bash is invoking... I'm not a wiley enough attacker, but I bet if you put something valuable behind it and put out a public challenge you'd be impressed with the opensshd vulnerabilities and other interesting features of ssh/tty/environment variables.
As you probably guessed, my comment was just listing a few of the things (today) that you're playing whack-a-mole with when it comes to making sure everything is secure. I think your "disprove me with code" approach is a great basis for a discussion.
Writing their own 'ssh server and shell' program is a different set of tradeoffs vs trying to fully lock down programs that are intended to be general-purpose... to be perfectly honest I don't think I'd want to be involved in either route, the stakes are stressfully high... but there sure is a lot less code to audit and test in writing your own ssh+shell than there is trying to secure and keep up-to-date sshd, pam, bash, etc. (especially since they'll be adding features that you explicitly do not want for your purposes)
I imagine the issue would also impact Electron, which might be the actual reason for blocking the format if they use that for some/all of their client apps (assuming the OP's claim is correct, I don't use it so can't check myself), because I'd like to think they could push out a back-end deployment quickly.
1: https://www.revenuecat.com/blog/growth/does-apple-keep-its-c...
2: https://www.techradar.com/news/how-steam-refunds-are-a-bless...
It's nothing to do with the monarchy - indeed, if anything I think the consensus amongst pro-monarchy folks is that the monarch is a constitutional escape hatch in the event of a government getting out of control.
Both the Conservative and Labour parties have an authoritarian streak to them when in government, and a solid percentage of the population seem to trust that our institutions know what they're doing and can keep their databases secure (although I hope that's been shaken in the last month with a number of police and electoral register database leaks).
Yes, although the article we're discussing shows that you can't rely on that, the dump could be subsequently moved to a developer machine for investigation, and unencrypted key material left in could be compromised that way... defense in depth would make sense here.
1: https://github.com/openssh/openssh-portable/blob/694150ad927...
1: https://www.youtube.com/watch?v=LTXI3zGkGtI (pretty sure he talks about the antimicrobial properties in this video, if not here then it's a similar video)
I think it's in our nature as software people to look at their ability to work with code, but they're quite good when applied to general language tasks. I've been using them for summarisation and reading comprehension and they're quite effective. I've also been working with a teacher friend on seeing if they can generate well-scoring essays on highschool English essays (as always, the problem is prompting and context).
On code, GPT-3.5 (moreso GPT-4) seems to have a good ability to generate and translate smaller scale code problems (hundreds of lines in low-boilerplate languages) but yes, they're like an eternally junior engineer whose work you're constantly having to oversee for subtle bugs, and I don't know that it actually saves time.
I'm pretty sure people are working on different approaches to applying them to code, with better prompting+context from larger codebases, and multi-step processing (i.e. rather than just a single prompt->response, letting the model iterate through a few steps independently, possibly guided by other adversarial/supervisor agent instances, testcase generators, etc.)
>If you're going to include the whole animal on the human side, you need to include the whole supply chain on the LLM side. The cost of building all the fabs and doing all the R&D to develop and manufacture model training-specific computers (matrix multiplier hardware). Just like with crypto, these resources had to be diverted away from other things (e.g. causing the price of gamers' graphics cards to skyrocket). It's only fair to interrogate the ROI.
That's fair, although it gets complicated to work out numbers because we don't train many LLMs, whereas we're constantly training humans, each of whom cost the planet tons of CO2 emissions every year... and, of course, your point that LLMs just aren't very good yet. I fear that they're good enough (or appear to be to the layperson) that execs will replace customer support staff with them, even if the outcomes overall aren't as good.
For that setup cost, you get a single-threaded human, starting to specialise in a single field. They will work in that field for around 7.5 hours every weekday, with ~20 days of holiday a year, for around 45 years, and then retire.
Llama v2 70B cost 1,720,320 GPU-hours[1] at 400W, so 688MWh. Once trained, it can be run 24/7, and you can spin up as many instances as you want on much lower spec hardware. That model produces output faster than a human while consuming around ~30W on my Macbook Pro.
Now, I know we're in pretty shaky spherical cow territory here, comparing a human (albeit a highly educated one) to Llama v2 in logical reasoning... but consider that this is the state of the art in generally-available LLMs after a few decades of research into machine learning & we're using repurposed silicon to compute vs the amazing complexity and physics-leveraging approach of human neurons... and the training cost is only 1 order of magnitude off humans.
Again, I'm not disagreeing with the general point of the OP, I'm not saying that the models we're using right now are the best/right ones (or that the hardware they're running on are the most efficient way of executing them), but I don't think the energy efficiency gap is actually all that high considering
Edit: If you look at the total footprint (considering not just the efficiency of the neurons, but the whole animal) the figures are very close - the llama model card indicates that v2 70B caused the emission of ~300T of CO2, and an average human in the US emits 16T of CO2 a year, so a human would emit ~320T of CO2 in 20 years. I assume children don't have as high a CO2 output, but even so it seems like it's the same order of magnitude.
1: https://github.com/facebookresearch/llama/blob/main/MODEL_CA...
Don't get me wrong, I'm not claiming that Machine Learning is as generally capable as animals/humans, and I don't entirely disagree with the OP (i.e. I don't know if our current approach has a chance of scaling to human-level capabilities). I just don't think computation-wise it compares that badly to animals, considering that it's the result of a few decades of work, largely on repurposed silicon
Do they really? They're certainly more energy-efficient in business-as-usual mode, but a human brain has 86 billion neurons, 600+ trillion synapses(!), and each instance takes 15-20+ years to train to do complex logical tasks. Even if the per-cell work is tiny (and, is it? cells are amazingly complex), 86 billion (or 600+ trillion) times 20 years is a lot of computation.