Depends on the model/provider. If you’re running this in Azure you can use their content filtering which includes jailbreak and prompt exfiltration protection. Otherwise you can strap some heuristics in front or utilize a smaller specialized model that looks at the incoming prompts.
With stronger models like GPT4 that will adhere to every instruction of the system prompt you can harden it pretty well with instructions alone, GPT3.5 not so much.
I’ve implemented a few of these and that’s about the most lazy implementation possible. That system prompt must be 4 words and a crayon drawing. No jailbreak protection, no conversation alignment, no blocking of conversation atypical requests? Amateur hour, but I bet someone got paid.
Let’s just hope for Gabe to live a long life still. Valve is a private company and not nearly as much in danger for enshittification as a public company would be.
But no, not like that. Clearly they need to find a way of materializing money from thin air.
The AI agrees with you.
Isn’t everyone looking at it for 2 seconds anyway before slapping on a case?
Alexas are all that, but also dumb as a rock. With ChatGPT im conversation mode I can talk aviut anything, Alexa falls on its dumb face almost immediately. Hope they inject some gen AI into their home assistants Stat!
Love(d) this movie. People I watched it in the cinema with back in the day thought it was boring. I thought it was amazing.
Mobile induction charging battery packs, I use one every single day and the only way those make sense is with magsafe.
Here are some more candidates:
I mean, we’re talking curl here, I don’t think a lot of suffering of consequences is happening. And man pages are often also not a great resource, throw everything at you, often don’t contain examples. If I’m building an app that integrates curl or libcurl, oh yes I’m reading the doc. If I just need curl to do something quickly, the LLM output is perfectly fine.
I recently watched the 3.5-hour workshop Mastering the curl command line by Daniel Stenberg, the author of curl
I commend the author for watching and subsequently summarizing this into a blog article, but in the age of LLMs like ChatGPT I cannot be arsed anymore to memorize any of the arcane command line arguments of curl, ffmpeg and the like.
BSH are the best. Respond super well to clicker training, wicked smaht and will never leave your side. Only caveat, they loathe being picked up and aren’t exactly a lapcat. Quiet and polite, bedt cat breed ever imo.
I think burkinis are dumb, personally, but I don’t give a damn what people wear. This is just typical ingroup-outgroup postering fueled by a proto fascist government, standard stuff really. Also, hygiene? Bro, it’s the ocean, a layer of clothing extra ain’t gonna make a difference.
Same, I was very confused what kind of burka cat toilet contraption we’re looking at here.
If there’s nothing good to watch, whatcha pirating then?
I use it because of inertia. I jumped from Firefox to Chrome back when Chrome was new and came with hot features, like tab isolation, or being ultra fast. Google has gone through enshitification though, and FF has gotten much better since those days. I’ve started the migration back to Firefox now though.
I don’t know, ads were always dog shit. Yeah you had your static banners and what not, but I remember the popup wars from the 90s.
I just asked GPT-4:
What’s 7 * 8 divided by 10, to the power of 3?
Its reply:
Let’s break this down step by step:
First, multiply 7 and 8 to get 56.
Then, divide 56 by 10 to get 5.6.
Finally, raise 5.6 to the power of 3 (5.6 * 5.6 * 5.6) to get 175.616.
So, 7 * 8 divided by 10, to the power of 3 equals 175.616
Eh, that’s not quite true. There is a general alignment tax, meaning aligning the LLM during RLHF lobotomizes it some, but we’re talking about usecase specific bots, e.g. for customer support for specific properties/brands/websites. In those cases, locking them down to specific conversations and topics still gives them a lot of leeway, and their understanding of what the user wants and the ways it can respond are still very good.