E. Alessandro

Into AI, design, and conversation.

[E. Alessandro @AIGuyAle]

@ComfyUI @higgsfield_ai I stopped using that shit of a platform a while ago, now I use @sumistudio_ai or comfy for local generations.

[E. Alessandro @AIGuyAle]

@sumistudio_ai Sumi is the goat

[Sumi Studio @sumistudio_ai]

New to the Sumi Studio canvas? Here's how it works: adding nodes, connecting shots, fixing one scene without touching the rest, and editing the final cut, all in one place. https://t.co/pAdEYru5Wh

[Nicky_42 @NickyLi42]

🧟‍♀️ MJ's Thriller, redone with 9 AI-model girls: DeepSeek, Claude, GPT, Grok, Gemini, Kimi, GLM, MiniMax and Qwen. For Halloween 🎃

Made with Codex + @sumistudio_ai. Characters by @ZipZipPipe (CC BY-NC-SA 4.0). This video is shared under the same license.

132s, all 54 original shots, original timing. How it was made below 👇

[Nicky_42 @NickyLi42]

Genie 4 just dropped (testers only). Honest take: I'm bearish on world models. Huge compute, meh UX, no use at scale yet.

But Gemini directing the world between inputs is basically a DM. So I mocked up a D&D table. Made in @sumistudio_ai, not Genie footage. Sound on 👇

[E. Alessandro @AIGuyAle]

@MANISH1027512 Damn great result and hot 😍
I am already using sumi, and loving it!

[E. Alessandro @AIGuyAle]

@NickyLi42 amazing result!

[Nicky_42 @NickyLi42]

Okay, Claude is the real deal. 🤯

I handed Claude Opus 5.5 our Sumi Studio MCP and it directed this whole 45-second Halloween motion piece by itself 🎃 12 scenes, 197 sound cues, every hit on the beat.

This one workflow could crank out brand videos and explainers all day, and spin up new accounts on other platforms fast. @claudeai you cooked. Sound on 👇

[BLOCKLORE @BLOCKLORDS]

Before any video gets made, someone has to write the script.

19 language models are coming to BLOCKLORE.

Claude: Claude Opus 5.5, Claude Sonnet 5.5, Claude Fable 5.1, Claude Haiku 4.5, Claude Opus 4.5, Claude Sonnet 4.5.

GPT: GPT 6 Luna, GPT 6 Sol, GPT 6.1 Sol, GPT 6 Astra.

Gemini: Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini Omni Flash.

Grok: Grok 4.7 for coding, agentic tasks and knowledge work, and Grok Build 0.1, a fast agentic coding model.

GLM: GLM 5.3, a 1M-token flagship that's strong at coding, agent tasks and cybersecurity, and GLM 5.3 Flash, an open-weight MIT-licensed model that takes text, image and video input

Qwen: Qwen 3.8 Flash, fast and low-cost with hybrid reasoning and image understanding. Qwen 3.7 Flash, the previous generation, stays on the list too.

Coming soon on BLOCKLORE.

[E. Alessandro @AIGuyAle]

@BLOCKLORDS Lrds 😌

[E. Alessandro @AIGuyAle]

@faucet_trade send ETH tokens to 0x7A2C69687D66549b1cC05b717ed7A08533F04adc
#base_sepolia_eth #eth #testnet #faucet #base_sepolia
https://t.co/1CCg639jrJ
Get more tokens at 👉 https://t.co/fZtDUPPaS4

[E. Alessandro @AIGuyAle]

@BLOCKLORDS Interesting results...

[BLOCKLORE @BLOCKLORDS]

GPT Image 2 vs GPT Image 2.5 Flare vs GPT Image 2.5 Sunburst:

We ran the same prompts through all three models to see what the 2.5 upgrade actually changes.

Here are the results.

Speed:

Flare is the fastest by a wide margin. It finished about 2.3x faster than GPT Image 2 and about 1.8x faster than Sunburst. Sunburst was still around 1.3x faster than GPT Image 2.

Image quality:

Sunburst came out on top, with more natural lighting and slightly fewer AI hallucinations. Flare's quality is close to GPT Image 2, with a big speed advantage.

Editing:

Editing is where 2.5 is supposed to improve the most, so we gave each model the same image and asked for several edits in a row. All three models handled the first edits well and kept the rest of the image consistent.

The differences showed up in the final edit:

• GPT Image 2 had a clear placement error.
• Flare had a small blending issue.
• Sunburst got everything right.

Pricing:

Both 2.5 models cost the same per token as GPT Image 2.

Which one to use:
• Flare for fast results and lots of variations.

• Sunburst for the best lighting and for images that need several rounds of edits.

[BLOCKLORE @BLOCKLORDS]

With the Agent Editor on Blocklore, just type your instructions and it handles the cuts, transitions, text, and effects automatically:

[Nicky_42 @NickyLi42]

People ask us what AI video costs per second. I wanted to look at what happens after you hit generate—and why a good-looking shot might still end up in the bin.
First in The Wrong Question. 🪐 https://t.co/DlaaXPse0D

[E. Alessandro @AIGuyAle]

@BLOCKLORDS @VitalikButerin Vitalik's Spanish is so good!

[BLOCKLORE @BLOCKLORDS]

This actually happened - just ask @VitalikButerin

[E. Alessandro @AIGuyAle]

@BLOCKLORDS Really good! Would love to see longer videos or short series

[BLOCKLORE @BLOCKLORDS]

Launch pricing is live on BLOCKLORE.

Seedance 2.0 Fast and Mini are discounted, and select credit packs are cut in the Credit Store, including flash deals while stock lasts.

The cheapest way yet to test what the new models can do.

[BLOCKLORE @BLOCKLORDS]

Coming soon to BLOCKLORE.

Seedance 2.5 enables our users to create breathtaking videos with longer cuts.

This video generation model is a massive leap forward for digital creators, filmmakers, and marketers.

And we can't wait to see what you create!