AI Insider / Tracking the race to AGI Startups • Agents • Open Source Sharing what matters before it goes mainstream

DM for Collabs
Joined August 2024
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
👀 cursor just put out a kit so anyone can build their own ai helpers you write what it should do in plain text, then run it on their servers or your own it even checks whether each change made your helper better or worse my guess: cursor wants to be where you build your ai workers, not just where you write code
The Agent SDK is now available. Define agents as code, with prompts in Markdown and tools in TypeScript. Deploy them on our hosted infrastructure, or run them on your own. Built-in evals let you measure every change and make your agents better over time.
1
Replying to @rubenssoto_ai
yeah you don't need a genius to click buttons and fill forms lol
5
Replying to @aymanemsi
yeah good enough for daily use is all most people need honestly
4
Replying to @vikktorrrre
anthropic gonna have to answer this one quick
8
Replying to @kimmonismus
feels like it's made for companies running it all day, not for us picking one model to chat with
28
NINE TIMES CHEAPER per task than Claude Opus 5.5. That's what OpenAI is claiming for GPT-6.1 Terra. - about 3x fewer tokens per task at max effort - a quarter of the price per token - reportedly quicker too OpenAI itself still tells you to use its bigger models for hard coding. I think this one is aimed at everyday stuff like clicking around apps and filling in forms for you. Cheaper and good enough, or the best no matter the cost?
openai just launched an ai that clicks around your computer for you and it's reportedly about half the price of the last one my guess: most people stop paying for the big models for normal work cheap and fast matters more when it's doing 50 small clicks a day would you actually let it use your apps without watching it? 👀
3
2
11
607
Replying to @IndependentEco
😂 what's the first thing you're letting it do
1
1
16
Replying to @ValencianaAbel
good and cheap is a hard combo to beat tbh
1
11
Replying to @ai_for_success
wonder if anyone's gonna run the same test on other open models now
31
Replying to @theaidesk_
Let’s wait for Astra
1
1
19
Replying to @Angaisb_
it's the computer use one tho, I expected it to be boring lol
1
201
Replying to @vikktorrrre
it's out now, $1 in $5 out so literally half of sol
56
Replying to @blueemi99
fair on efficiency
1
111
Replying to @IndependentEco
sonnet feels like the fair fight, they're both the cheaper everyday ones
1
1
55
Replying to @arena @OpenAI
gonna wait for these scores before i believe the computer use hype
111
Replying to @LuminaBench
WTF is benchmark ?😭
172
Replying to @ValencianaAbel
half of sol's price and close on most benchmarks, sol users gonna feel it
1
2
41
openai just launched an ai that clicks around your computer for you and it's reportedly about half the price of the last one my guess: most people stop paying for the big models for normal work cheap and fast matters more when it's doing 50 small clicks a day would you actually let it use your apps without watching it? 👀
OpenAI
@OpenAI57m
GPT-6.1 Terra: delivering fast and capable computer use at a lower price. GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.
3
1
14
1,009
Replying to @OpenAI
wait so this is on plus too? figured the screen control stuff would stay pro only for a while
4
2,755
Replying to @LuminaBench
Sure, if it's actually good at that price though i'm not complaining
1
49