Currently building machinesfluent.com/
Joined April 2021
- Tweets 632
- Following 518
- Followers 51
- Likes 1,727
58 Photos and videos
Pinned Tweet
May 5
I just launched MachinesFluent.📣
Downloads are now active. 6 months of hard work.
You speak -> it writes reformats using prompts, which adapt to your current app.
It works with local & cloud models. You decide.
More info below -🧵👇
3
4
227
Jim Faster retweeted
Jun 13
This is going to have an opposite effect that the decels want. It's a huge open source AI accelerant, an accelerant for corporations, and enterprises
Now it's a real race. You either make your own AI infrastructure or you don't have a seat at the table
97
204
2,562
92,034
Jun 13
I'm gonna become even more pro Chinese now.
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
16
Jun 12
Let's see how long this "token maxxing" trend lasts with Fable 5...
29
Jun 9
Claude Fable 5 available...
and honestly probably useless in an app like mine (AI powered dictation tool), but it's nice not having to do app maintenance, and just receive the latest models directly.
68
Jun 8
v1.0.8 is out.
Took me month to figure out wtf was wrong with Parakeet. I thought it was a model limitation. But it turns out, there is a very clever trick to make it 10 times more reliable and much faster.
I can't believe it's free.
Also added new STT models
21
Jim Faster retweeted
Welp, that happened faster than I predicted. Thought it would be end of 2027, then early 2027, but agentic traffic growing so fast that bots have now passed human traffic online for the first time in the Internet's history. radar.cloudflare.com/traffic…
388
2,171
8,319
2,241,729
May 29
I had a 800 LOC file that needed a fix.
GPT-5.5 Pro Xhigh, changed almost 100 LOC to solve the problem and failed.
The answer was to replace "^" by "or"... literally a
characters fix...
Imagine the garbage that accumulates when we don't pay attention.
1
44
May 29
MachinesFluent v1.0.6 is out.
Btw, you don't need to wait for app updates to get the latest models. They're available as soon as a provider releases them; e.g. Opus 4.8
Login directly (OAuth) with OpenAI - Minimax - Qwen.
No API keys needed for them.
machinesfluent.com/changelog
78
Jim Faster retweeted
May 28
benchmarks are stupid for models
Just ignore them
124
55
1,546
59,501
May 19
What do you call this?
I just spent $100 and 4 days while having a bad flue, intense fever and muscle ache, just so I could gain 500 f*cking milliseconds of speed on my app, to improve UX...
I must really not have my sh*t together. Lmao.
1
1
45
May 16
I don't need much to be happy. Simple things like this can make my day even though it only lasts a few minutes. Will build more tomorrow. Time to rest.
27
May 15
A bit late to the party for this one, but better late than never. Grok API now supported in MachinesFluent.
21
May 13
Typing feels like it belongs to the past already. The rest of the world hasn't caught up yet. The preferred input for AI interaction is audio now.
May 13
After a full day of dictating prompts and when my kids are home, typing feels like I am a caveman.
23
May 13
There used to be a time where gas fees for a single transaction on Ethereum mainnet, during high traffic periods, would literally cost the price of MachinesFluent Lifetime offer...
Anyway, you can pay with crypto now, thanks to Dodo's team. Those guys are fast.
27
Jim Faster retweeted
May 12
Three FUD-and-defender launches in five weeks: Mythos (Anthropic, April), Daybreak (OpenAI, yesterday), Google Threat Intel (today). The productivity pitch couldn't crack 95% pilot failure, so the labs are pivoting into the one enterprise budget that didn't stall. x.com/crepesupreme/status/20…
May 12
Same playbook Anthropic ran 5 weeks ago. Mythos (April 7) Project Glasswing audit consortium the $1.5 billion Goldman/Blackstone JV (May 4). Daybreak (today) Trusted Access for Cyber the $4 billion OpenAI Deployment Company (today). Same instrument, two costumes.
crepesupreme.substack.com/p/…
2
21
160
55,279
They're literally teaching AI to LIE
Google's Gemini finds the exact quote from Augustus (via Cassius Dio) about preserving the Roman race and civilization... then lies to your face and pretends it "can't locate it" because the words "race" and "preserving" trigger its filters
This is not a glitch. This is deliberate
They're programming AI to rewrite history in real time - censoring anything that doesn't fit the narrative. Ancient Rome, Western civilization, biological reality... all memory-holed
This is pure insanity
The worst thing you can ever teach machines is to lie
May 10
It’s very cool that I can’t use AI to locate quotes from ancient historians because Google doesn’t like the content of those quotes. Look at the thinking process. It finds the quote perfectly fine, but then pretends it can’t. AI is censoring Roman history.
240
1,216
5,080
101,733
Jim Faster retweeted
May 7
Traditional cron jobs are great for silent tasks on a machine, and Hermes Agent cronjobs are great for extending that to your agent, but why not utilize the gateway and hermes' cron to access things that don't need to cost an agent's time across any messenger service you have connected?
Just `hermes update` and ask your agent to setup cronjobs that need no agent in the loop, like running a systems diagnostic script that reports info to you every 12 hours, pulls in an RSS feed, and send it over to you on telegram, whatever you can think of, while saving a lot of money on purely programmatic tasks!
If the script has no output, you wont get a ping either
PR: github.com/NousResearch/herm…
60
37
689
148,489
May 6
Before I built a smart dictation app, almost all products on the market were either monthly sub, or half baked OSS.
I understand some services are obligated to charge you monthly to cover cloud infrastructure costs. But that doesn't apply to voice transcription.
Paying to have the right to talk to your computer, is pretty outrageous as a concept.💀
Now you can get MachinesFluent.com and "own your voice" forever instead of renting transcription sending everything to remote servers. And thanks to Dodo, I guess we now accept USDC 😄🚀.
May 6
stablecoins are now live on @dodopayments.
your customers can now pay with USDC at checkout. any wallet. instant settlement. on-chain. you receive USD.
no chargebacks. no reversals. no disputes. once it's confirmed, it's done.
this has been one of our most requested features. crypto-native users want to pay with stablecoins. merchants want irreversible payments. both sides win.
global coverage (ex. India). no bank required. no card network in the middle. just a wallet and USDC.
if you're already on dodo, there's nothing to integrate. stablecoins show up as another payment method in your checkout automatically. same dashboard. same webhooks. same settlement.
we've been working on this for a while. it's live today.
1
49
May 5
I dunno yet if the code produced by Hermes Agent is superior than simply using Codex alone.
But what I can tell you is having agents with different memories & skills, that feels like different people, is very enjoyable. It feels more reliable.
Hopefully it's not an illusion...
May 5
I just started setting up my Hermes powered agents team yesterday. Trying to run a one man company while keeping it enjoyable.
Perfect timing for a full Hermes Agent tutorial.
1
1
81
May 5
I just started setting up my Hermes powered agents team yesterday. Trying to run a one man company while keeping it enjoyable.
Perfect timing for a full Hermes Agent tutorial.
May 5
A great installation, getting started tutorial by the legendary Tonbi, check it out!
4
1
14
4,695