Currently with:
- Claude Pro
- Cursor Pro
I find Claude Pro/Code the best overall but am getting frustrated with token usage at the moment.
Currently with:
I find Claude Pro/Code the best overall but am getting frustrated with token usage at the moment.
----
Creator of whatsthesalary.com and whatstheincometax.com
|
|
|
Since my use is only casual, I stick to free services like ChatGPT. I have a lot of curiosity and it is fun to be able to ask questions as they pop into my mind. My AI experience with diagnosing technical issues has been horrible, though. It is like a dumpster dive into rabbit hole city. If the choice is between changing a single setting and low level reprogramming the entire app, the AI will always go for the latter.
Plesse igmore amd axxept applogies in adbance fir anu typos
I mix my usage between free tiers of Claude, ChatGPT, and Perplexity - though Perplexity is what I use for less important things, it isn't as effective for my queries. I also have an Amazon Q developer professional subscription from work, which is pretty effective.
If I was going to pay for AI it would probably be Claude. It's less sycophantic and gives me more useful answers than the others.
OpenAI API for receipt and invoice processing via Paperless-NGX and Paperless-GPT.
All other duck.ai for privacy
The little things make the biggest difference.
I have no interest in an online Hackenthorpe Book of Lies.
I use Office 365 Personal without CoPilot, and I disable AI options whenever I can, eg disabling smart features in Gmail.
If I wanted hallucinations I would use drugs.
No one has come up with why I should put my trust in any of them.
Right now .... none.
I'm researching getting a locally hosted voice assistant going using Home Assistant and Ollama. But early stages
Mike
Claude Pro, but I've also found the recent token consumption changes frustrating.
I have an RTX 5070 Ti GPU in my main desktop & access to a DGX Spark device with 128GB shared memory for local AI work & development. If I didn't have these, I would have likely blown through their cost many times over if I used subscription services instead.
I use Github copilot pro for all coding related tasks.
Then I use a mix of gpt, claude and perplexity free tiers for all other queries.
I got a free sub to Perplexity - and I like it.
It's amazingly good as an information aggregator.
Somewhat like having a team of researchers.
I don't use it to make decisions - but to get me information so that I can make decisions.
Given that it gives its references - I can double-check them.
None
I can’t think of a reason why I would these days.
Software Engineer
(the practice of real science, engineering and management)
A.I. (Automation rebranded)
Gender Neutral
(a person who believes in equality and who does not believe in/use stereotypes. Examples such as gender, binary, nonbinary, male/female etc.)
...they/their/them...
None for personal use. All free ones for testing and experimentation.
You need to know everything about how something works if you want to avoid it for yourself and others.
- NET: FTTH & VDSL, OPNsense, 10G backbone, GWN APs
- 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
roobarb:
No one has come up with why I should put my trust in any of them.
Of course you shouldn't trust them. They're excellent for research, search, development, product suggestions, but you need to independently verify anything that comes out of an AI from a reputable source if you're using it for anything important. It's a massive time saver in many cases.
timmmay:
Of course you shouldn't trust them. They're excellent for research, search, development, product suggestions, but you need to independently verify anything that comes out of an AI from a reputable source if you're using it for anything important. It's a massive time saver in many cases.
It can also be a huge time waster. Important to know what to disregard.
Plesse igmore amd axxept applogies in adbance fir anu typos
Rikkitic:
It can also be a huge time waster. Important to know what to disregard.
Genuinely interested in an example of a time-wasting instance.
Can you recollect one and rough out how it either misled / fooled you ?
Here's a concrete AI example - from today:
I'm looking at renting a car in Europe for September, for an extended drive in some Eastern block countries which (a) I've never been to & (b) have a more reasonable cost-of-living than western Europe. eg: Poland, Romania & Bulgaria.
OK: I'm considering two vehicles, each of which has pros & cons. A significant bonus would be to be able to fold down the seats and sleep in the back (in a pinch) when I can't find an AirBnB and it's raining too hard to put up the tent.
20 years ago, all car brochures had nice detailed drawings of the dimensions (inside & out) - but now that's boring and you are lucky to get the overall outside length & width. Note that neither of these models is sold in NZ - so I can't just go & look.
I'm aware that somewhere out there on the interwebs, some car nut(s) will have published this data in a forum - but where and in what language ?
I'm quite good at Search Engine / key word searches - but this question is too complex. It's a bunch of general words combining in many possible ways - to mean a very specific thing.
Perplexity can find it for me - in 5 seconds - and give me the reference. Then I can go and read it and decide if it's credible.
The old route - which I also did yesterday (to ask about Nav maps in Euro vehicles & their geo-political boundaries) - is to join one of the 5 Renault car forums in the UK, get registered & beg for help. In a week I may attract some answers - although most forum members are driving cars that are pre-Nav... so I may not get anything.
Honestly ? AI is like Internet .2
pdh:
Genuinely interested in an example of a time-wasting instance.
Can you recollect one and rough out how it either misled / fooled you ?
Not how the AI's seem to get the series number wrong. This is a series 9, 10kg dryer...
Perplexity "Best":
Perplexity "Claude Sonnet 4.6":

Google
|
|
|