One half of @ctpt_studio. Building music systems for wellbeing and movement.
Joined March 2008
- Tweets 35,230
- Following 6,200
- Followers 9,716
- Likes 15,157
4,859 Photos and videos
Pinned Tweet
Apr 28
We’re announcing something new today: Motif - music that moves with your yoga class.
Instead of static playlists, the music follows the structure and energy of a class and adapts on the fly.
3
1
10
740
Tero Parviainen retweeted
Introducing the Fusion API, the smartest compound model in the market.
Fusion achieves Fable-level intelligence at half the price.
How it works 👇
514
1,326
11,548
3,649,534
Tero Parviainen retweeted
Jun 13
The new beta Claude Code model switcher
28
302
5,076
170,593
Tero Parviainen retweeted
sorry what was that about EU regulators
2
6
563
Tero Parviainen retweeted
Jun 12
one of the underrated harms of AI is the death of vernacular design
now the kebab shop sign is gonna look like claudeslop instead of chaotic dafont goodness
37
317
6,043
123,437
Tero Parviainen retweeted
Jun 12
mfers realizing that an AGI company is a competitor of every company on earth
it's only a matter of time
Jun 12
Exclusive: Anthropic’s rise to nearly $50 billion in annualized revenue is giving it new leverage over customers.
Some startups now worry the AI company is reserving its best technology for products that compete directly with them.
Full story: thein.fo/4utqJpN
15
28
1,019
162,524
Tero Parviainen retweeted
Jun 12
hi there! we couldn’t help but overhear you discussing biology from across the room. we really appreciate your enthusiasm, but we’re going to need that conversation to come to an end now
83
219
4,725
205,532
Tero Parviainen retweeted
Jun 11
Added a loop sequencer ♻
You can share your loops via URL.
sascacci.com/touchwaves/?seq…
#touchwaves #webaudio #audiovisual
2
19
94
4,391
Tero Parviainen retweeted
Jun 11
Meet LiveBand: a real-time AI jamming companion! 🎸
It generates live music accompaniments with zero perceived latency ⚡️
It runs locally on Macbooks, can generate any instrument (more than one at a time), is wildly robust, and is trained from scratch on a single GPU! 🧵👇
8
13
93
4,633
Jun 11
Claude Fable one-shotted* a spring reverb for me today.
Wrote the algorithm, tuned it against a reference impulse response, using a Nelder-Mead optimiser over the parameter space. Also generated these plots, then inspected them.
*with four rounds of listening evals by me
5
1
53
2,496
Jun 11
I didn't ask it to do any of this specifically. I just wanted the reverb. Using this type of self-initiated backpressure, instead of just trying to wing it, is very good model behaviour.
1
5
380
Tero Parviainen retweeted
Jun 11
Holy crap! whosampled.com/sample/541728…
Jun 10
Learned that ZZ Top sample Autechre on their 1999 album
1
3
12
1,380
Tero Parviainen retweeted
Jun 10
Easy solution to slow down recursive AI self improvement:
- The lab with the top-ranked model must agree THEY must not use it for working on frontier AI
- But everyone else should have access to it.
By definition, this means the frontier doesn't advance.
51
67
838
69,234