![]() ![]() ![]() |
|
MadEngineer:
Were there any distribution or release specific issues with its answers?
I gave him the exact versions. He gave me CLI instructions that actually led to the goal. It was a PHP versioning problem with incompatible package dependencies.
- NET: FTTH, OPNsense, 10G backbone, GWN APs, ipPBX
- SRV: 12 RU HA server cluster, 0.1 PB storage on premise
- IoT: thread, zigbee, tasmota, BidCoS, LoRa, WX suite, IR
- 3D: two 3D printers, 3D scanner, CNC router, laser cutter
An Ai , or perhaps more correctly artificial 'simulation' of intelligence does...
Human things that make conversation flow along?
Its quite human to infer, guess or create an answer from whatever is closest.
Maybe a commercial release will come with a confidence ranking?
Much like your Tesla might decide that if it stopped for every truck it thought 'might' be there its not going to get you anywhere.
https://www.thedrive.com/news/33789/autopilot-blamed-for-teslas-crash-into-overturned-truck
https://insideevs.com/news/494714/tesla-crashes-with-semi-investigation-underway/
https://arstechnica.com/cars/2019/05/feds-autopilot-was-active-during-deadly-march-tesla-crash/
Though now it seems that some Tesla drivers are getting spooked by phantom random braking, so don't follow one too close.
A hard working Ai just can't win.
https://www.autobodynews.com/tesla-phantom-braking-class-action-lawsuit-includes-model-3-model-y.html
So ChatGPT does the human, inference, guess, or just make stuff up to satisfy you with an answer.
It's not good at saying it can't give you an accurate answer.
Apparently reviews of its mathematics results have identified problems.
So these people below looking at pairing it to pass those details to a Computational program.
First look - ChatGPT + WolframAlpha (GPT-3.5 and Wolfram|Alpha via LangChain by James Weaver)
https://www.youtube.com/watch?v=wYGbY811oMo
Dilbert, 16 Jan 2023
Dilbert, 17 Jan 2023
Here we go, excerpt from the HA homepage:
„… we’re introducing a new rule for the Home Assistant community: it’s no longer allowed to use ChatGPT or other AI systems to help others.
Although these systems generate elaborate and well structured answers, they are wrong. Often they are wrong in subtle ways which only someone with the right expertise could detect. And those people wouldn’t need AI to have written that answer.
Using an AI system to generate an answer while not providing attribution to AI will result in a ban. If you use attribution, we will delete your post and issue a warning. This also means suggesting someone “ask ChatGPT” is not an acceptable response.“
- NET: FTTH, OPNsense, 10G backbone, GWN APs, ipPBX
- SRV: 12 RU HA server cluster, 0.1 PB storage on premise
- IoT: thread, zigbee, tasmota, BidCoS, LoRa, WX suite, IR
- 3D: two 3D printers, 3D scanner, CNC router, laser cutter
I can understand that. How long before it inadvertently generates some malicious code that someone blindly tries to use?
Seems we will see less 'alternative facts' from ChatGPT.
New Job category in the making Ai fact checker :-)
Oh another update, pity it does not come with fact checking or confidence level mark.
Microsoft are probably close to releasing its own SuperBing with ChatGPT
Though how inserting advertisers products and messages in replies will be interesting.
Then there is Google etc and their response.
neb: Some commentary on ChatGPT (linked rather than inlined to avoid copyright issues): Dilbert, 16 Jan 2023 Dilbert, 17 Jan 2023
Thank you for picking out some decent Dilbert for me. I have really soured to it since Scott Adams became a Trump shill.
But those strips were pretty bang on with ChatGPT I think.
Jarle Dahl Bergersen | Referral Links: Want $50 off when you join Octopus Energy? Use this referral code
Are you happy with what you get from Geekzone? Please consider supporting us by making a donation or subscribing.
Generated Python code to provision a Azure Digital Twin and the corresponding Python code to send data, as a simulated IoT device
The little things make the biggest difference.
Bing is getting AI
Introducing your copilot for the web: AI-powered Bing and Microsoft Edge - YouTube
Yo Bing, give me a mean plan with x requirements.
Yo Bing, give me a shopping list for the above mean plan organised by grocery section
Tom Scott has put out a really interesting discussion about how ChatGPT and similar might in hindsight be as disruptive as Napster back in the day.
Get your business seen overseas - Nexus Translations
ChatGPT lied to me last night.
I asked it (is that the right pronoun?) which Sony cameras come with an FTP client, within the response it included the A6600. This is blatantly wrong. When I responded to point this out it replied that I was indeed correct, but no actual acknowledgment it had given the wrong answer initially.
Did Eric Clapton really think she looked wonderful...or was it after the 15th outfit she tried on and he just wanted to get to the party and get a drink?
floydbloke:
ChatGPT lied to me last night.
I asked it (is that the right pronoun?) which Sony cameras come with an FTP client, within the response it included the A6600. This is blatantly wrong. When I responded to point this out it replied that I was indeed correct, but no actual acknowledgment it had given the wrong answer initially.
How does one access Chat GPT? Is there an app?
Handsome Dan Has Spoken.
Handsome Dan needs to stop adding three dots to every sentence...
Handsome Dan does not currently have a side hustle as the mascot for Yale
*Gladly accepting donations...
It works in a browser. Just got to https://openai.com/blog/chatgpt/ and sign up.
Get your business seen overseas - Nexus Translations
|
![]() ![]() ![]() |