[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 152260807275i6.png (247 KB, 798x448)
247 KB
247 KB PNG
I know a lot of people will cry scream, and shit themselves with rage about it. But it's happening and it's unstoppable. Just because you used GPT-3 3 years ago and it was shit doesn't mean the cutting edge of AI today is bad. Most models have improved leaps and bounds since you last used them. Or you used them for stupid reasons like telling you a poem instead of using them to tack extremely difficult mathematical and coding problems, or hard science and engineering problems. AI is as good as you prompt it, and if you prompt it stupid it gives you mediocrity.

Even if AI capabilities had a hard freeze at their current capability. And only incremental refinements in hallucination reduction and common sense were done - they would still change the world. But we're not anywhere near the ceiling.

The way models have been scaling up this early into the compute build up makes it undeniable. AI is already, right now, smarter than 98% of people. Stupid people are just not able to tell when they're speaking with a genius.
>>
Then it runs out of context and goes retarded.
>>
>>107543766
How many r's in your post?
>>
>>107543822
37
>>
File: 1628786074242.png (451 KB, 496x496)
451 KB
451 KB PNG
>>107543766
it was obvious from the start
>>
>>107543797
A lot of people wouldn't be able to answer what they ate for dinner two days ago. Cutting edge AI has a context window Its short term memory) as long as a thick textbook. And it's only a matter of time until short term memory is a solved problem, and incorporation of new data into long term memory is a solved problem.
>>
>>107543766
lol wasnt it already this smart back in like 2008-2012
>>
>>107543861
>A lot of people wouldn't be able to answer what they ate for dinner two days ago
Because it is meaningless information and the brain evolved for efficiency, not perfect memory.
>>
>>107543876
A lot of people will ask AI stupid questions so most models will need meaningless information
>>
>>107543861
Short term memory is already a solved problem if you use tools to allow the AI to dump its relevant context into a text file for later retrieval.
>>
Turn on Claude Plays Pokemon on twitch. This piece of shit cant even beat pokemon red.

>but its likely going to replace my job anyway
>>
File: b.jpg (41 KB, 640x285)
41 KB
41 KB JPG
>>107543766
will never pass my riddle test
>>
>>107543766
Those random text generators are now even more retarded than 2 years ago. What are you on about? 2 years ago chatGTP built functional software with very little user input, but it has since been rendered complete retarded. You used to be able to actually vibe code with very little effort, but they completely massacred and lobotomized it.
>>
>>107543830
We've cum a long way
>>
File: 1743436049665323.png (2.03 MB, 1080x1920)
2.03 MB
2.03 MB PNG
>>107543766
Imagine AGI or even ASI!
I was just reading about how SkyNET used to skin humans alive for their skin for the T-700's. Fun!



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.