If the company controls the enclave, and controls the software that contacts the enclave, how do you know that your messages only flow to the enclave and aren't also logged somewhere else?
Attestation can get you the idea that the software they're running is what they say they're running, but unless you constantly audit every version, how do you know that it operates exactly how they tell you?
I mean how do you know there's really an enclave there at all :D If your threat model is "this entire provider is untrusted and potentially hostile" it seems like there's lots of options for them to do bad things, that can't be fully solved by secure enclaves.
I definitely wouldn't trust standard Linux style containers that expose a shared Linux kernel at the moment, there's been far too many LPE and container breakout vulnerabilities this year. It's possible that in future if the kernel gets a lot more hardened, that could change but things like Firecracker are a better bet from a security standpoint.
The historical precedent is those people get other jobs, and/or everyone else in the economy gets richer faster than they get poorer.
If this does or doesn't happen with AI kinda depends how fast it makes multiple different sections of the population unemployed, rather than just one group like "software developer" or "vehicle driver".
That's very much not what happened in the Industrial Revolution. Hand weavers were relatively well-paid, and the new factories were more likely to employ the urban poor - including kids - than to give jobs to people who used to be hand weavers.
It took maybe a century and a half for the modern consumer economy to appear. And it didn't happen until money started being shared into the wider economy for political reasons, not technological ones.
Meanwhile the enclosure of common land was catastrophic for rural populations. Sources of food and pasture disappeared as the Big Houses took them over and turned them into huge single family estates.
Your claims are mostly orthogonal to mine; I'm not saying "this was fun for all involved" or "the specific people who were laid off got better jobs" or anything like that, though with the benefit of hindsight I can see how:
> those people get other jobs, and/or everyone else in the economy gets richer faster than they get poorer.
Could be read the way you seem to have read it. With this quotation "other jobs" was "a job" rather than "a good job", while "everyone else" was intended to refer to the aggregate economy, not to all other individuals within it personally and individually benefitting. And the "and/or" because this isn't just about the initial phase of the industrial revolution but all technological job displacement that has continued to this day.
Not even saying it was fast or error-free; another comment I made around the same time used a colourful metaphor to reference various economic depressions that keep on happening.
> The historical precedent is those people get other jobs
Heh, got fired from my old company for reasons not connected to AI. Then couldn't find job. AI suggested one that is perfect for me, helped me put up webpage (I can do everything from electronics/firmware/backend/iot/cad, but nice webpage is not something I can do easily), then helps me every day to learn new things and make useful products I always wanted to make. And I found something safe for now from AI and big companies - making old equipment work with new systems. No one will touch it, because you have only small projects here and can't extract a lot of value at once, but a single person who can do anything with some help from AI can carve out a whole niche.
I think such solution - thousands of small niches by small entrepreneurs (1 person companies) who can use AI for some small skill they are lacking - are our future and they will absolutely extract a lot of money from people, who hope to make some more money with small optimisations of already working businesses. One example of such small optimisation - making a slightly better wood drying controller for small woodmill owner. He already put up small cabinet with cheap tuya controllers and needs to control them though atrocious app. Now I can make a small module that will add several more sensors, give him some simple dedicated control panel on webpage. Not even medium company would make such small product, because planning phase alone would cost more than what I can charge for whole system and do it in a week.
And it all started with one prompt - "Given my CV, what single person job would you recommend?"
One other option, if you're an obsidian user is, just add the excalidraw plugin there, and then you can ask your agent to just create excalidraw diagrams as needed.
Definitely Opus 5.5 via claude code has no problems generating Excalidraw images with no additional software, just based on conversation.
For some cybersecurity tasks, the Chinese models are already good enough, things like PoC development or things like exploiting mis-configurations.
Whilst I'm sure the top-end OpenAI/Anthropic models might be better, I've found their guardrails so twitchy (especially Anthropic) that I wouldn't try to use them for even vaguely security related work.
In case it's useful, I did a talk about giving talks earlier this year which covered various things I've encountered/learned over the talks I've done and seen
https://tagt.mccune.org.uk/ - has sections and links to specific parts of the video, so you don't have to watch the whole thing to get to a bit that seems interesting :)
Jokes can be tricky in talks, as they rely on shared cultural context. If you know your audience well and know what will land with them it's great, if you don't then care is needed.
I generally advise people to put some jokes in if they like but don't make the talk rely on people getting the jokes.
but don't make the talk rely on people getting the jokes
Adding to this, don't automatically include in your joke delivery. Make the joke - if it lands and the audience is laughing, you can pause until it dies down. If the joke doesn't land, you just move on and everyone will forget about it.
This seems like a very nice release. Just ran it over my Kubernetes security benchmark that I run for most new releases. It was fast, cheap, and got a high scoring result, nice!
It's great that they're making the auto usage tokens free by default and I guess auto mode will be a good default for a lot of workloads, but recent changes to the auto mode classifier just moved me to either use YOLO mode or use a different harness.
I've been using Opus 4.6 for some security related work (it has much looser guardails that later opus models) and last week, all of a sudden, the processes started to fail. It wasn't the main model blocking commands but the auto mode classifier changed how it worked and it started blocking the main models commands.
That's one specific incident, but it does have a wider potential problem which is, if you use Anthropic's harness you'll always be at the risk of sudden breakage from server-side changes that are opaque to the end user, which is a tricky one for building long lasting processes.
I had the same issue with Codex, where one day it just became unusable.
Weirdly, this would occur with very common commands like git push, where the classifier would start to yap about not being able to verify the repository's privacy settings.
My bug reports went ignored, the issue continued, and I eventually just switched to Full access.
I have not yet encountered an issue with Claude's Auto mode.
What works in that case is adding a message saying "I authorise you to do $thing", which the classifier treats as explicit consent to do something and execute commands toward that goal.
Same Story as it ever was. The first time I encountered what I thought was a phishing attack at the bank I worked at 25 years ago, it turned out to be a marketing campaign, with URLs that put our company name as a user before the domain name (back in the day when creds could go in the URL).
Attestation can get you the idea that the software they're running is what they say they're running, but unless you constantly audit every version, how do you know that it operates exactly how they tell you?
I mean how do you know there's really an enclave there at all :D If your threat model is "this entire provider is untrusted and potentially hostile" it seems like there's lots of options for them to do bad things, that can't be fully solved by secure enclaves.
reply