@devteamdrew

creating interesting things in code

devteamdrew.com/
Joined August 2017
  • Tweets2.6K
  • Following470
  • Followers2.8K
411Media
DreW retweeted
DreW
@devteamdrewOct 1
I asked Claude Opus 5.5 to create a family portrait All style and direction was produced solely by Opus @claudeai
3
3
59
3,953
DreW retweeted
DreW
@devteamdrewOct 2
I asked Fable 5.5 to create its own introduction video This entire video was created with just a single prompt It created everything from the visuals to the music without any further instruction by me @claudeai
6
6
3
122
9,622
DreW retweeted
DreW
@devteamdrewSep 27
I asked @claudeai Opus 5.5 if it could make itself on a cymatics plate 9 attempts, 1 physics bug, 6 tones later: 97.9% of the sand formed Claude. Every step in the film is real. Repo in the comments👇
5
5
30
14,916
|￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣| | Don't Push To Production On Friday | |＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿| \ (•◡•) / \ / —— | | |_ |_
82
370
57
2,434
83,113
Verification of the facts is still needed. But imagine how useful this will be in education! Teachers are going to have a great time trying out new fun ways to create resources.
DreW
@devteamdrewSep 15
I set @claudeai Opus 5 xHigh on a /goal of creating an infographic of the water cycle as a test of its visual intelligence and video creation capabilities Everything in this video was created by Claude It took just over 4 hours to go from 0 to a final MP4 render of over 4 minutes Here's what it created:
1
1
357
Replying to @devteamdrew @claudeai
Really gorgeous. Nice work!
3
1
57
8,485
GLM-5.3 on my NYC test. It rendered the city black. GLM-5.3 has no vision. Text in, text out. I described the black screen to it and gave it a chance to fix it. It couldnt. It cant see what it made. In 2026 a coding model that cant see its own output is not shipping visual work. Zhipu says its the best open weights coding model and it beats Fable 5 and GPT-5.6 Sol on three benchmarks. On my prompt it produced 2,005 lines and a silhouette. Grok 4.6, Gemini 3.7 Flash and Opus 5 can all see their own render. Thats the whole difference.
20
6
2
154
26,826
Cut through the noise
1
39
Replying to @justalexoki
Decentralized, open source, version control platform & tool with git compatibility. xitter.cf/bsmntech/status/208328… github.com/tankrap/keel & github.com/tankrap/hull
bsmnt.tech
@bsmntechJul 31
What we have been working on

The future of source control

Article · @bsmntech
1
167
New in Claude Code: your sessions can now message each other. Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.
1,903
4,604
2,475
59,911
7,981,162
DreW retweeted
OpenCode
@opencodeAug 2
DeepSeek Flash did 8T tokens on August 1st 5T of free usage + 3T on OpenCode Go
192
216
103
5,821
993,368
Some big things happening here at bsmntech. We've decided to integrate into another product to bring a seamless integration of git, operations, and development in one platform. Stay tuned :)
3
5
123
Replying to @bridgemindai
Prediction for bro's next post: "FABEL IS THE BEST, I'M VIBE CODING WITH IT ALL DAY, BUT I DON'T TRUST ANTHROPIC."
1
1
207
created this entire brand identity using Fable 5
32
18
3
617
25,741
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication. We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating. 🧵👇
660
2,055
1,374
14,018
6,016,408
DreW retweeted
DreW
@devteamdrewJun 11
Lemonade Voice is officially live. A brand new voice application built to completely change how you dictate and build. It is 100% FREE to download No more paying for subscriptions like WisperFlow Run it locally or connect a cloud API key. Hook up Groq for free access to Whisper v3 and v3 Turbo. Agents can dynamically edit your dictionary and snippets via skills. Download it directly from the site: getlemonade.dev/voice FREE Hands-free building starts today. 🎙️🍋

Lemonade Voice — Talk to Type in Any App

Free speech-to-text for your desktop. Press a hotkey, speak, and your words land as text in any app. Local-first whisper.cpp or bring your own API key.

getlemonade.dev
1
1
11
1,650
DreW retweeted
Froots
@frootsaiJun 3
Froots: - Infinite agents - Infinite databases - Agent browser - skill library - routines that connect to widgets - internal kb - internal memory - no additional downloads on install Hermes agent desktop: - cli wrapper - browser and memory are external features that download on install - memory is external - no built in Knowledgebase
2
3
11
321
Multitasking vs focus
196
811
76
5,844
1,082,316
DreW retweeted
DreW
@devteamdrewMay 26
New model soon then
Margins
@themarginguyMay 26
We are seeing a sustained statistically significant degradation in Claude Code with Opus 4.7 since last Friday May 22nd at marginlab.ai/trackers/claude…
1
1
1
296
Try Composer 2.5
BridgeMind
@bridgemindaiMay 20
New CursorBench results just dropped. Two big takeaways. Composer 2.5 is way better than most people think. 63.2% score at $0.55 per task. Nearly matching Opus 4.7 Max and GPT 5.5 Extra High at 20x less cost. This is insane value. Gemini 3.5 Flash is #10 at 49.8%. Below GPT 5.5 Low. Below Opus 4.7 Low. Google's newest model can't even beat budget tier competition. Composer 2.5 is the sleeper. Gemini 3.5 Flash is the disappointment.
1,953
1,873
133
19,242
8,152,024
Load more