Bard seems to still make mistakes more often than chatgpt. A couple of pretty strange ones in this answer.
I appreciate you. Could you appreciate me with a simple Like? 🖤
the quality is baffling, they’re fumbling the bag pretty hard imo
Yeah, it’s almost like you need both. Bard more has more current data to tap into, but OpenAI’s data is more reliable.
one friend of mine has observed that recently Bard has gotten much worse, right as sama says chatgpt has been fixed https://x.com/nptacek/status/1754206370813874395
But bard is very good in summarising ling YouTube videos. Did you tried this?
What’s up Vitya, how are you doing? What will happen on 7 February?
Have you played around with Grok? How would you compare it? I find myself gravitating to it more than gpt given its access to X data + less reminders / disclaimers of it being an AI and hesitating to answer questions.
but chatgpt has gotten slow af, i literally have to refresh the tab everytime before using it
We can’t trust the solutions given by ChatGPT or Bard. We have to validate the output 😧
There have been numerous instances where Bard made blunders, particularly when it comes to dates and years.
Does that mean ChatGPT is still the best right now?
The best thing about Bard is that it has access to Google's Knowledge Graph, so it's good at finding, connecting, and summarizing recent information, and pointing you directly at relevant blog posts or documentation. Bard's reasoning can get weird, but IMO it's better than ChatGPT for navigating cutting edge research.
Last time I compared these two, Bard also was much more censored and attempting to reeducate you
Please read this pdf. It is phenomenal https://shorturl.at/eltUW " it is not sufficient to understand the opportunities that exist for groups or individuals to actualize their needs. It is necessary to analyze to what extent the environment represses, tolerates or stimulates opportunities. "
It makes simple mistakes, but Bard seems a bit better to prototype. Bard gives newer code some other LLMs, spew often old unoptimized implementations.
as a native chinese speaker: the answers seems generally in the right direction. although some parts like the Japan bit was off
Not even close to chatgpt. Even the lower version of chatgpt will provide you with a legit answer.
I'd imagine you were a bit young for Everquest during its heyday but I think it might have been the most imperfect class in that game. https://eqoa.fandom.com/wiki/Bard
There’s this notion that any well-capitalized company with nearly unlimited resources can create an accurate and useful AI. I’ve never believed this to be true. There’s a finite amount of top-tier talent that will draw towards the power law winner in this AI arms race.
I tried to use Bard when researching cloud computing companies for a Business Dev role I had. It was brutal, context was bad but the actual response held no validity when further researched. It would pull fake answers and provide links to back them up that had no relevance.
Bard feels like a year or two away from ChatGPT. From experience, I’m having trouble making it understand priming prompts. It is also having difficulty stringing together additional prompts with the original prompt.
I've been paying for Google Ads on that topic specifically to mislead Bard. Sorry about that.
You got a feed your learning model that fine vintage copywritten intellectual property and not pay the owners of the content for it… Snack on Grade A hallucinations, without any of that ethical baggage of paying the rights holders for it.
