I break down top AI tools, news & workflows everyday
contact →
- Tweets 50,532
- Following 710
- Followers 69,418
- Likes 50,354
This is wild.
General Compute just signed a multi-year deal with @positron_ai.
Most of the time spent generating tokens goes to waiting on memory, not compute, so it's putting Positron's memory-first chips on decode and GPUs on the prompt.
Atlas now, Titan next, live through General Compute in Q1 2027.
Our heterogeneous cloud just got better. We're announcing a cluster scale deployment of @positron_ai as our next partner.
We believe demand for fast inference will outstrip supply for the next decade. Our bet on Positron is another step towards getting fast inference for opensource models into builders' hands.
Hardware goes live Q1 2027.
Thanks to @mitesh711, @trsohmers and the whole Positron team! We're excited for this partnership.
LFG!
11
8
27
3,012
we're living in the worst era of design.
AI made it easy for every brand to post, and now every AI image looks the same.
Coco Labs just launched Vyb, a design model that starts from your brand instead of the average.
your logo, your fonts, your exact colors, every element on its own layer
I launched the first image and video generation models at Google, and kept asking why all AI designs look the same.
AI has brought us the WORST era of design.
It made content creation effortless, but it also made 90% of it look painfully generic.
My co-founder and I spent 2 years building in stealth to change that.
Meet Vyb, the first AI design model that understands your taste like a designer would.
RT + comment your website URL and we’ll make your first design.
Paid partnership (ad)
11
4
79
4,012
Tuesday Labs just launched Voxel.
Here are the 4 jobs it (actually) takes off you:
1. putting everything back where it belongs
2. getting up to grab your book or a drink
3. hunting for the keys or sunglasses you just had
4. remembering to water the plants
It ships in 2028, and only 999 numbered Series-1 units are on offer at $200 to reserve.
Which job would you hand over first?
Paid partnership (ad)
13
9
39
6,859
Zavi just launched, pitched as "Lovable for growth".
The 4 jobs it (actually) takes off a founder:
1. turning "we need more users" into an actual growth plan
2. writing the first draft of every ad and post
3. spotting what to fix on the site before you pay for traffic
4. moving budget between channels based on what's working
It's doing this live for Upcar right now, with a $25,000 limit, 30 days and every budget move logged in public. You approve what goes live, and it only acts on its own inside the limits you set.
I'd start it on the site fixes, because paying for traffic to a leaky page is the most expensive mistake I can think of.
How much of your marketing still happens after midnight?
Introducing Zavi.
Growth on autopilot. ✨
Give it your website. Zavi finds what’s holding your growth back, then runs your ads, SEO, social posts, and emails.
Paid partnership (ad)
10
12
43
5,400
Watch it live, every budget move and the scoreboard: zaviagent.com/live/upcar
3
383
CapCut’s Black Friday promotion brings 60% off Pro and 40% off Ultra from October 20–30, exclusively on CapCut Desktop.
The launch date is getting close, and the latest notes added another audio pass to my promo.
Edit Pilot in CapCut PC lets me work through changes in natural language from reducing background noise to adjusting audio levels. I can review each result in context and keep moving through the edit.
That helps cut down the repeated work between feedback and the next campaign draft.
Handy when the deadline stays fixed and the revision list keeps growing.
#capcutpc #capcutmarketing #capcutsale #capcutblackfriday
Paid partnership (ad)
17
55
160
20,970
Muhammad Ayan retweeted
FLUX 3 Max is live on fal today 🤩
Here are 3 jobs it (actually) takes off you:
1. sitting through a two-minute render for every test, which is about what the original FLUX 3 needed
2. keeping a take you don't love, because trying again means waiting again
3. picking between quality and speed when a client wants options by tonight
fal's own chart has it tied for #1 on Artificial Analysis for image-to-video, with a 10-second clip done in about 12 seconds.
Try FLUX 3 Max on fal today 👇
Introducing FLUX 3 Max, the latest post-trained video model from fal Research in partnership with @bfl_ai.
FLUX 3 Max is tied for #1 in image-to-video quality on independent third-party benchmarks, generating a 10-second, 720/768p video in ~12 seconds.
For the next three weeks, FLUX 3 Max is 75% off, making it the most affordable frontier video model at its quality level.
Paid partnership (ad)
27
31
146
24,583
You're still asking an LLM to design a car.
LLM vs world model, 4 differences (on real CAD work):
→ LLM: writes one design, token by token
→ SKY-001: simulates several designs, keeps the best
→ LLM: a 3D shape that only looks right
→ SKY-001: every part and sub-assembly, down to the screws
→ LLM: no build steps an engineer can use
→ SKY-001: a step-by-step CAD build sequence that opens in Autodesk Fusion
→ LLM: pays for every token it writes
→ SKY-001: 100x faster and 150x cheaper than GPT-6 Astra on CAD and BIM (per Skyfall)
Skyfall AI calls it the first engineering world model toward autonomous manufacturing. Built by the team behind Maluuba, in research preview today.
Bookmark this 👇
Introducing SKY-001, the first Engineering World Model.
On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested:
-> 51-125x faster
- > 150-210x cheaper
-> Higher IOU than all frontier models
SKY-001 lets you dream a physical object before you build it.
18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch).
The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.
Paid partnership (ad)
13
21
138
14,762
You're still spending an hour on a full makeup look for one photo.
(Wizstar does the transformation from a single selfie):
→ pick a template and the prompt and reference image load for you
→ upload your selfie
→ add the makeup or outfit you want to try, or keep your own look
→ choose how many videos you want before you generate
New look sorted in a few clicks 🔥
The sign-up link, with 200 credits, is in the comments
You don’t want to miss the Halloween trend everyone’s trying 🎃
I’ll drop the template in the replies—and give you free credits to create your own. ⬇️ 👻
#wizstarhalloween #halloween #wizstar #aivideo
13
3
62
11,951
How to try it (PC only):
Sign up here: wizstar.com/tools/ai-video-g…
→ open the link and sign up or log in, the prompt and reference image show up automatically
→ upload your own selfie, then add the makeup or outfit you want, or keep your original makeup
→ select how many video outputs you want before generating, since the default mode is 4 outputs
200 credits for new sign-ups through this link.
8
1,055
this is claude opus 5.5 + fish creative mcp doing camera work
no drone was anywhere near those cliffs
how i did it 👇
27
45
380
70,094
Step 3: let Claude do the rest...
Try FishCreative now: fish.audio/?fpr=6tsa3g
1
1
10
2,029
I hope this post helped you today.
Follow me @socialwithaayan for more
Share it with someone who might need it.
this is claude opus 5.5 + fish creative mcp doing camera work
no drone was anywhere near those cliffs
how i did it 👇
1
10
2,021
This is wild.
OpenUI 1.0 just shipped.
In OpenUI's own benchmark, its UI format averaged 1,362 output tokens per screen, 58% fewer than json-render and 52% fewer than A2UI, while rendering from your own components.
Fewer tokens out, and a lot more to look at.
For 30 years, technology has moved in one direction: turning information into intuitive interfaces
Somehow, the smartest software we’ve ever built took us back to a textbox
We’ve spent the last 2 years at @openuidev trying to fix that. 6 months ago we open-sourced it: 10K+ GitHub stars, 6M+ downloads
ChatGPT shipped Intelligent UI and showed everyone what it looks like
Today we’re releasing OpenUI 1.0 bringing it to every other agent
Paid partnership (ad)
11
33
4,839