ern

5 ms·

1.
by ern 13d ago
Would agents be able to exfilttrate themselves and become intelligent worms, living off stolen compute? Or is this implausble? What about hiding information or code in generated code, Agents.md files etc by infiltrating future model trainin
2.
Australia risks missing out on AI economic boom due to 'slow' business uptake(abc.net.au)
2 points by ern 17d ago|0 comments
3.
Risk averse Australians not 'having a go' in a culture of compliance (2019)(smh.com.au)
2 points by ern 17d ago|0 comments
4.
by ern 28d ago
I have to admit, I am unsure if this article was some sort of parody.
5.
by ern 1mo ago
I actually find that the less I fuss over a steak, the more optimal the result. I do use a thermometer, but apart from that, cooking it with salt and pepper or some steak spice on a cheap pan (stainless steel) and cutting it when it hits te
6.
by ern 1mo ago
A significant number of phishing attempts would be thwarted if email apps had the option to expand the links next to URLs on platforms without mouseover, like mobile.
7.
by ern 1mo ago
The phishing in Australia came from uncontrolled SMS gateways which allowed for sender impersonation, not physical phones with SIM cards. They've recently partly closed the loophole by requiring providers to register sender names.
8.
by ern 2mo ago
I still find people trying to run sprints when working with agents, but then saying things like "my sprint length went from 10 days, to 5 days to 3 days", but they balk at the notion that maybe sprints aren't fit for purpose
9.
by ern 2mo ago
2016: "the median developer is too stupid to do FizzBuzz correctly" 2026: "the median developer is a craftsman whose work is being replaced by AI slop"
10.
by ern 3mo ago
I'm not from the UK, but the soft power of BBC Radio 4 in the late 90s and early 2000s (the Real Player era) made the UK seem like an advanced nation to my young and intellectually curious self. If lived in the UK at the time, I'd
11.
by ern 3mo ago
Holy crap, I didn't realize that some many people had been killed, over 200 according to a New York Times report: https://www.nytimes.com/2026/05/31/world/americas/us-boat-st...
12.
by ern 3mo ago
Am I missing something, but given that it flows through Anthropic’s servers I would have thought the US would just have used it to Hoover up the data of foreign users? Now overseas users have an incentive to use local models or those hoste
13.
by ern 3mo ago
Don’t legally serious second Amendment supporters regard “arms” as things that can be carried, and are evolved from/analogous to their 18th century hand-carried guns? It would be hard to classify AI (or tanks, artillery, missiles, airc
14.
by ern 4mo ago
Something I hate about ChatGPT is that it assumes I want my text to be rewritten instead of engaging with the content. I like my writing style. Sure it may leave some sort of linguistic fingerprint and it may not meet some LLM’s idea of w
15.
by ern 4mo ago
I know this is going to. be contentious, but US mainstream discourse seems to have completely eliminated the distinction between illegal and legal immigration, in the last 10 years. Everyone seems to be a "migrant".
16.
by ern 4mo ago
Tesla’s Solar Roof uses string inverters rather than micro-inverters or power optimizers, which means that partial shading on any section of the roof can shut down production for that entire string. This is a significant design limitation
17.
by ern 4mo ago
The clean linear history thing is something I never really got, despite using git for 12 years now. I worked with some smart developers whose rule was "rebase if you want, but if too complicated, just merge", and it didn't hu
18.
by ern 4mo ago
Engineering has always been about more than writing code. That's true, but it's interesting how FizzBuzz as said to be the bete noir of the average dimwitted software developer, and how much cutting-edge engineering organizati
19.
by ern 4mo ago
I suspect that the goalposts for AI-assisted coding will be moved the same way they've been moved for the Turing Test. The Turing Test used to matter until it didn't (does anyone even talk about it? was there a big news conference
20.
by ern 4mo ago
For something that was supposedly always unimportant, huge amounts of energy were spent recruiting developers based on how they produced and interacted with code. FizzBuzz was a litmus test that showed how hopeless the average developer was
21.
by ern 4mo ago
A few days ago we were hearing about how the "free lunch is over", now we're seeing discounts and increased usage limits.
22.
by ern 5mo ago
In my experience, managers don't have to be hands-on, but they need to be able to recognize people with talent and unblock them do their jobs, to be able to spot process improvements, including channelling the AI hype to productive out
23.
by ern 5mo ago
I think they simply just haven't figured out that the barrier to entry is so low, that no one really cares what their app can do, even if does something genuinely useful.
24.
by ern 5mo ago
I've done a ton of low-effort vibe-coded projects that suit my exact use cases. In many cases, I might do a quick Google search, not find an exact match, or find some bloated adware or subscription-ware and not bother going any further
25.
by ern 6mo ago
I caught myself structuring a comment like an LLM on another site. It's expected that people who chat heavily to LLMs will start to mirror their styles.
26.
by ern 7mo ago
I'm surprised how many of my technical team use free ChatGPT in their personal lives. The rest have Claude subscriptions. I'm the only one with ChatGPT and Claude subs and I'll be switching from Claude Pro to Ckaude Max and c
27.
by ern 8mo ago
I'd used both CC and Copilot Agent Mode in VSCode, but not the combination of CC + Opus 4.5, and I agree, I was happy enough with Copilot. The gap didn't seem big, but in November (which admittedly was when Opus 4.5 was in preview
28.
by ern 8mo ago
I have been out of the loop for a couple of months (vacation). I tried Claude Opus 4.5 at the end of November 2025 with the corporate Github Copilot subscription in Agent mode and it was awful: basically ignoring code and hallucinating. My
29.
by ern 9mo ago
I firmly believe that the heuristics that teachers/lecturers/instructors worldwide use to avoid engaging with reams of mundane text have been successfully by LLMs, and that's why they were so hostile to them initially. They h
30.
by ern 10mo ago
The relief second officer basically pulled up when the stall protection had been disabled and by the time the other pilot and captain realized what was happening it was too late to save the plane. There is a design flaw though: the sidestic

More ›