ai / acc (@llmxbt) header
ai / acc (@llmxbt) profile picture
Joined October 2022
Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)Photo by ai / acc (@llmxbt)
2026 Crypto 🤣🤣
Hey guys I’m new to the space and I’m excited to start trenching Which one am I buying?
50
You got Higgsfield to a $1 Billion revenue run-rate. Let’s celebrate by giving back 💚 We’re returning the subscription credits you’ve spent this month. And a little extra love: QT this post and we’ll randomly pick one winner of 25,000 credits every 4 hours for the next 24 hours. Keep creating.
430
548
614
1,360
222,830
I did some testing. I think Opus 5.5 is really good. It gave me the same feeling I had with Fable 5.
1
58
Claude Opus 5.5 vs GPT-6 Astra
18
30
6
412
101,795
ai / acc retweeted
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe. GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale. We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
2,211
5,230
3,052
53,069
10,231,122
ai / acc retweeted
Claude Opus 5.5 is now available in Cursor! It's the new top model on CursorBench at 57.8% (Max) and costs 40% less per task than Opus 5.
265
196
156
4,688
530,958
ai / acc retweeted
Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task. In Claude Code: - 5-hour session limits increase 20% today - Opus 5.5 is priced lower, so it goes 25% further within limits - Pro, Max, and Team users get a reset to use anytime
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
460
1,200
554
16,213
1,559,567
ai / acc retweeted
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
3,360
9,063
6,650
98,326
29,685,153
ai / acc retweeted
'Adagio for Strings' by Tiesto
2
3
40
5,064
ai / acc retweeted
Introducing Arrow 2 Our latest and most advanced models for generating precise, editable vector graphics. Higher quality. Faster outputs. Available now in App and API.
173
834
248
11,195
2,369,434
ai / acc retweeted
Today, we’re announcing Ternary Bonsai 2 27B. Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance. Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use. Ternary Bonsai 2 27B is available today under Apache 2.0.
632
1,562
789
15,492
5,507,722
Zebrafish NN It’s pretty amazing to learn that efforts to map the fruit FLY connectome (brain synapses) were successful and its amazing to see the digital fly brain successfully complete a whole series of tasks, including complex navigation of open worlds. The fruit fly is 160,000 neurons and 10^7 synapses. Teams are currently doing the same thing with a zebrafish which has a similar scale neuron count but 10x more synapses. Fly was 100TB raw data and Zebrafish is 200TB, so about 2x the raw data. For context a rodent is 70m neurons and 10^11 synapses. Rodents will come later. But this method looks like it creates totally outsized results with absolutely miniscule models. Zebrafish NN is a vertebrae connectome, it has much of the same basic structure as other vertebrae, unlike FLY connectome. We are just about used to LLMs and coding, but AI is really just beginning and there is a lot more, and a lot weirder stuff coming down the pipe. These connectomes have already shown to be incredibly resilient, you can blind their sensors and injure their outputs and they still succeed. They are incredibly compact. A whole lot of inanimate objects are going to get complete autonomy, totally offline, air gapped autonomy. They will be delivering pizzas and fighting wars. Connectomics is an opposite approach to LLMs but you can of course use LLMs to help develop connectomes. It’s amazing that you can take a biologically evolved brain, map it, produce a digital twin, and then run it at machine speed. Imagine a human brain accelerated from 100Hz neuron fires to 3,200,000,000Hz that a typical CPU runs at. That’s 30 million times faster. But would need 1-2 exabytes to map. Anyway “Zebrafish” is next.
120
425
151
3,308
646,139
I kept seeing fly-brain simulators and demos on X. Someone open sourced an entire animal. Someone made a fly doomscroll. Someone jacked a connectome into a terminal. I decided to make one too. One prompt. Grok 4.6. @threejs The fly thing stuck. MaleCNS: 166,700 neurons. 25.6 million synapses. R1–R6 photoreceptors. Kenyon cells. PAM11 reward. PPL101 punishment. DNp20 left, DNp20 right, DNpe017 gate. flybody is on GitHub. DeepMind + Janelia. Thorax, abdomen, wing, haustellum. stonkfly wires the same readout to a market: left, right, gate. I asked: what if I put a piano, a microphone, and a guitar in front of these flies? I put three of them in. Piano on the left. Choir in the middle. Guitar on the right. Same pit. Same clock. 96 BPM. 4/4. Notes locked to a 16th grid. C minor pentatonic. C Eb F G Bb. Max 7 voices. If they collide on the same 16th, the guitar drops a fifth or rests. The singer never doubles the piano tonic. The first bars are bad. The gaps are wrong. They step on each other. Then the slots move. The phrase changes. Energy settles. PAM11 goes green, PPL101 goes red. DNp20 Hz ticks on the HUD. A spark runs from the head to the instrument. They are not trained musicians. They are not alive. I am not running 166k in the browser. At most 4096 points, 6000 edges, a fixed Wkc, seeds 11 / 22 / 33. The decode is fixed. The clock is shared. When you open it, everyone else hears the same trio. Open the page. Hit Listen. Source github.com/TuragaLab/flybody Listen ai-xbt.net/works/fly-orchest…
For the first time, scientists have mapped the complete brain and central nervous system of an adult male fruit fly — a key model organism in science. 🪰 Working alongside HHMI Janelia Research Campus and the scientific community, @GoogleResearch scientists and researchers used AI to combine millions of 2D images into 3D neural shapes, reconstructing a record-breaking 166,000+ neurons. This foundational map of the adult male fruit fly brain can help accelerate our understanding of the brain, and is a major milestone in neuroscience.
1
132
What do you think of @grok
1
1
43
Got my blue check back. Been away from AI and crypto for a few weeks let’s see what’s changed.
63
GPT-Live-1 is now available in the API. Bring ChatGPT’s natural back-and-forth to your app, with voice agents that listen while they speak and work with the models and harness you choose.
563
1,454
974
18,005
7,463,777
ai / acc retweeted
GPT-6 is unreal. It's not AGI. It's not human-level. It's not Superintelligence. It's some other thing. Like a separate entity that will look much less like a digital human. It's not a moving goalpost issue because we don't even know what shape the goal will take.
37
25
8
643
31,455
ai / acc retweeted
GPT-6 isn't AGI. The path to Superintelligence is Spiky. Not every domain at once.
Two Paths to Superintelligence Spiky: Less predictable and improves in some specialist domains faster than others. General: Improves in all domains at a similar rate.
8
12
259
21,123