It promises users that the new Astra model (GPT-6) can do "anything you can do on a computer. Fast".

RIP to creatives, people in IT, and analysts who mostly work on their computer.

the cull is coming

huge news for the dumbest person you know

Fable 5.1 outsold

Sorry but in the promotional material they literally use this model for the most useless tasks imaginable.

1 minute ago, Kota said:

Fable 5.1 outsold

astra is eating her up in every benchmark, it's gonna end ha

my website development about to skyrocket

4 minutes ago, KatyPrismSpirit said:

Sorry but in the promotional material they literally use this model for the most useless tasks imaginable.

It's because that's what the target audience uses AI for…

I wish they showcased more useful use cases. Like developing code for apps, or using it for repetitive tasks in a Shopify store or something. Not "put this table on Ebay" or "Order food" thats not how AI should be used…

I wish my company would just move from Anthropic to OpenAI. Fable consumes too many damn tokens. I'm constantly reaching my 5 hour limit and I find Opus to be too full of sh*t.
My personal tests with Astra yesterday have all been very impressive.

1 minute ago, Badgalbriel said:

I wish my company would just move from Anthropic to OpenAI. Fable consumes too many damn tokens. I'm constantly reaching my 5 hour limit and I find Opus to be too full of sh*t.
My personal tests with Astra yesterday have all been very impressive.

they dont give you a license at work?

Just now, KatyPrismSpirit said:

I wish they showcased more useful use cases. Like developing code for apps, or using it for repetitive tasks in a Shopify store or something. Not "put this table on Ebay" or "Order food" thats not how AI should be used…

i get what you mean but they are obviously trying to appeal to the GP and show that it's for everyone, us developers and scientists already know its capabilities and use it, if they promoted it with a "developing code for apps" video the geepee would think it's for coders or something, they have to make it appealing for everyone's use cases as a business there is no "AI should only be used for this type of stuff" limits anymore i'm afraid...

4 minutes ago, Badgalbriel said:

I wish my company would just move from Anthropic to OpenAI. Fable consumes too many damn tokens. I'm constantly reaching my 5 hour limit and I find Opus to be too full of sh*t.
My personal tests with Astra yesterday have all been very impressive.

Fable is unusable for me mostly , and im on the 5x max plan . I'll use Fable to do the beginning prompt of a project to get it as fleshed out as possible on first shot ( which uses most if not all of my tokens ) and then switch to Opus 5

Edited September 4Sep 4 by May

4 minutes ago, Badgalbriel said:

I wish my company would just move from Anthropic to OpenAI. Fable consumes too many damn tokens. I'm constantly reaching my 5 hour limit and I find Opus to be too full of sh*t.
My personal tests with Astra yesterday have all been very impressive.

opus is autistic and talking to it makes my blood boil

but how did you get access to astra sis? it's not rolled out yet to the geepee, i'm perched to get it

4 minutes ago, SlimyGhoul said:

huge news for the dumbest person you know

I wanted to name a few members here but the ban isn't worth it.

OT: Ready for OpenAI to dumb it down a few months in

10 minutes ago, KatyPrismSpirit said:

they dont give you a license at work?

Yeah, I have a license for Anthropic. But even with my license, Fable eats all my tokens constantly.

8 minutes ago, RideOrDie said:

opus is autistic and talking to it makes my blood boil

but how did you get access to astra sis? it's not rolled out yet to the geepee, i'm perched to get it

Let's say I work for a company that's really close to both companies

Just now, RideOrDie said:

i get what you mean but they are obviously trying to appeal to the GP and show that it's for everyone, us developers and scientists already know its capabilities and use it, if they promoted it with a "developing code for apps" video the geepee would think it's for coders or something, they have to make it appealing for everyone's use cases as a business there is no "AI should only be used for this type of stuff" limits anymore i'm afraid...

true, there arent. but i think AI could be used in much more meaningful ways than "order food" or "put this table on ebay". they should feel the responsibility to actively promote people to seek the value from it. otherwise it's a waste of electricity, really.

If they want chatGPT to become more accessible to the public and for more people to use it, they should show what value it can bring. if someone who doesn't have a clue what AI is sees this, they will think 90% of the time "I can do these things myself already, I'm good".

If you roll out this new model, might as well show use cases how it can help people in truly complex and time consuming tasks. like managing finances, creating overview on their computers, optimizing their website to boost sales.

1 minute ago, Badgalbriel said:

Let's say I work for a company that's really close to both companies

let me in

I'm sorry, but I'm really not buying it. It largely feels like more nonsense overhyping and attempting to claim that we're "in the AGI era" or whatever. As a user of these technologies in a professional setting, literally none of this matters while agentic tasks still cannot be reliably replicated with at least the accuracy of other modern processes - while it is still non-deterministic in nature and can't be relied on for consistent results.

The promises of current agentic and generative AI technologies are honestly grossly inflated compared to real capabilities. By engaging with the doomer scenarios, people are falling for their marketing and directly contributing to how deeply ingrained generative AI is in our lives and how people and companies see it as a way to abandon workers and their livelihoods with no safety net in place.

Edited September 4Sep 4 by Cruel Summer

Every model they put out is apparently going to change the world, and then nothing happens.

Turning a clever little chatbot into something that creates genuine economic value for businesses, is incredibly difficult.

These tech companies will continue to say 'isn't this brilliant' but adoption will be slow, and uneven. I wouldn't worry too much.

5 minutes ago, CallumDavies said:

Every model they put out is apparently going to change the world, and then nothing happens.

Turning a clever little chatbot into something that creates genuine economic value for businesses, is incredibly difficult.

These tech companies will continue to say 'isn't this brilliant' but adoption will be slow, and uneven. I wouldn't worry too much.

yeah this is just giving siri with additional steps

Investorslop. Can confirm none of this **** actually works in real life use cases.

Graphic designers are really gonna be jobless very soon.

These always look cool in their presentations but never work in practice. I am yet to find an LLM that is correctly able to list all Madonna songs that had featuring artists on them. Literally all of them, free or paid, fail to execute this seemingly simple task

I stopped using AI for a lot of my work because I do it so much better and faster. I still test models maybe once per month on the same task, waiting for the day it can actually make what I need it to. OpenAI and Anthropic like to claim frontier models are good enough, but they don't understand visual quality and consistency are more important to creatives than the speed at which it can order salad for me on UberEats.

Edited September 4Sep 4 by Archetype