AssemblyAI ✓
@AssemblyAI · joined 20 Oct 2017
Access powerful AI models to transcribe and understand speech via a simple API. Try our no-code playground for free 👉 https://t.co/YPCK9mqDG6
45 759Followers
414Following
2 760Posts total
111.1KViews on collected posts
Latest posts
Looking for new posts…
@AssemblyAI Dictation is easy to Block. "You Boost, I Block."
8 views · 0 likes
· 0 reposts · 0 replies
Open on X →
@AssemblyAI Exactly. Making dictation feel instant and affordable at scale is where the real engineering challenge lies.
18 views · 0 likes
· 0 reposts · 0 replies
Open on X →
Dictation is easy to prototype and hard to ship. Making the speech-to-text → cleanup loop feel instant, at a price that survives production volume, is the actual work.
Our forward-deployed engineers walk through it live, Wed Sept 9. Plus, get $20 in credits to build it https://t
24.2K views · 7 likes
· 1 reposts · 2 replies
Open on X →
@dan_aai builds voice agents at AssemblyAI, which means he thinks about how they're actually put together more than most people do.
His framing: a voice agent isn't one model—it's two roles.
🔹 The Responder holds the call almost the entire time—voice activity detection,
1.9K views · 13 likes
· 2 reposts · 9 replies
Open on X →
If you're building dictation: our forward-deployed engineering team is hosting a workshop on building and optimizing dictation features on Wednesday, September 9—live examples, the full pipeline, Q&A throughout.
Register: https://t.co/uRTmXSHRNj https://t.co/dQvwBGHTnw
1.6K views · 6 likes
· 2 reposts · 2 replies
Open on X →
Qwen3.5 4B is now on AssemblyAI's LLM Gateway—optimized and hosted by our team for the fast rewrite tasks at the center of voice products: dictation cleanup, transcript rewrite, live formatting.
On voice rewrite tasks it averages 612ms—1.9× faster than GPT-4.1, at 94% lower cost
11.4K views · 9 likes
· 1 reposts · 6 replies
Open on X →
@AssemblyAI @ycombinator @MollySOShea @sourceryy The lawn-care chain automating back office work with Lovable and Claude Code is the real market shift here. Once the buyer is ops instead of engineering, the bottleneck becomes handoff and auditability. I would package approval sta
3 views · 0 likes
· 0 reposts · 0 replies
Open on X →
9/
His read on the next 18 months: get it to work, then figure out the honesty.
On-device models, sovereign deployments, and voice as a dimension added to everything—not a screen you shout at.
Watch the full episode 👇
https://t.co/hV5HszQkqa
360 views · 1 likes
· 0 reposts · 0 replies
Open on X →
8/
Which leads to the question he thinks the industry is dodging: should a voice agent tell you it's AI?
Text agents do. Voice agents try to pass as human because the moment they disclose, people hang up.
@YouveGotFox thinks that has to change.
281 views · 1 likes
· 0 reposts · 2 replies
Open on X →
6/
And plenty's still broken. Stand 3 people next to a humanoid robot and it can't tell who's talking to it.
Same over the phone: background speech derails the agent.
Humans can close our eyes and still follow a conversation. Voice models mostly can't—yet.
43 views · 1 likes
· 0 reposts · 1 replies
Open on X →
5/
The models grew into it. The newest models take context about their environment, including our latest Universal-3.5 Pro.
A drive-thru AI used to have no clue it was taking a McDonald's order—now you can tell it to lock onto the customer and tune out the kids yelling in back.
49 views · 1 likes
· 0 reposts · 1 replies
Open on X →
4/
Then the buyer changed.
It's an API, so the customer was always an engineering team. Now a lawn-care chain automates its back office with Lovable and Claude Code—on voice infra.
Since coding agents have blown up, "everyone's an engineer now." TAM up ~100x.
54 views · 1 likes
· 0 reposts · 1 replies
Open on X →
3/
How much? Weekly conversations through AssemblyAI's API are up 800% in three years.
A peak week now runs 120M+ conversations and 2M+ hours of audio — ~4x YouTube's daily volume.
~100M API calls a day, 1M+ developers.
138 views · 1 likes
· 0 reposts · 1 replies
Open on X →
2/
His framing: for the first time ever, voice is a reliable form of data capture.
Not just "AI voices", but can a machine actually understand you, at scale, in the real world?
That has shifted in the last ~2 years, and the volume followed.
522 views · 4 likes
· 0 reposts · 5 replies
Open on X →
1/
In 2017, Dylan Fox joined @ycombinator's first-ever AI batch. The perk: $100K in GPU credits.
His bet: voice was terrible, but it would get good—and when it did, developers would build everything on it.
This week with @MollySOShea on @sourceryy, he showed the receipts 🧵
70.6K views · 14 likes
· 2 reposts · 6 replies
Open on X →
Against accounts of the same size
15 posts from the last 90 days, next to the 10K–100K follower range. below its peers on both reach and engagement.
Median views281this account982median for 10K–100K
Reach, %0.61%this account3.52%median for 10K–100K
Engagement, %0.64%this account1.92%median for 10K–100K
| Metric | This account | Median for 10K–100K | Ratio |
|---|---|---|---|
| Median views per post | 281 | 982 | 0.29× |
| Reach (views ÷ followers) | 0.61% | 3.52% | 0.17× |
| Engagement rate | 0.64% | 1.92% | 0.33× |
Others in this range → Compare with another account → How these benchmarks are built →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
52231 Jul
138
54
49
43
281
360
3
11.4K26 Aug
1.6K27 Aug
1.9K28 Aug
24.2K4 Sep
18
85 Sep
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
1.72%31 Jul
1.45%
3.70%
4.08%
4.65%
1.07%
0.28%
0.00%
0.15%26 Aug
0.64%27 Aug
1.35%28 Aug
0.05%4 Sep
0.00%
0.00%5 Sep
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.92%.
What the audience does
Likes51.8%59 in total
Reposts7.0%8 in total
Replies31.6%36 in total
Quotes2.6%3 in total
Bookmarks7.0%8 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Similar accounts
The 11th Hour with Ali Velshi ✓
@11thhour
91.5K followers
· 12K posts
Irish🇮🇪✝️ ✓
@CapatanEl4662
91.5K followers
· 5.6K posts
Prefectura Naval Argentina
@PrefecturaNaval
91.5K followers
· 6.9K posts
Darren ☘🇮🇪 ✓
@CryptoIRELAND1
91.5K followers
· 95.4K posts
Blake Neff ✓
@BlakeSNeff
91.5K followers
· 2.6K posts
DFlow ✓
@dflow
91.5K followers
· 1.9K posts
Elie Merheb ✓
@_ElieMerheb
91.5K followers
· 52K posts
Gen-Z for Change ✓
@genzforchange
91.5K followers
· 3K posts