
Building @OpenGradient @Tribeca_xyz | Techno-optimist views
New York City
Joined June 2023
- Tweets1.6K
- Following5.3K
- Followers3.5K
- Likes5.6K
Matthew retweeted
Seedance 2.5 Flash is now in the DreamFort Video Studio.
It's very fast and costs the same as Seedance 2.0. Find it under Balanced in the model picker.
Try it: dreamfort.ai/video
Matthew retweeted
DreamFort is private and uncensored AI. Chat with abliterated models, make images and videos, or put yourself or friends in trending scenes with ReCast, and none of it is tied to who you are. Watch the whole video.
dreamfort.ai
Private and uncensored AI
Private and unrestricted intelligence, available at @dreamfort_ai
DreamFort is private and uncensored AI. Chat with abliterated models, make images and videos, or put yourself or friends in trending scenes with ReCast, and none of it is tied to who you are. Watch the whole video.
dreamfort.ai
Private and uncensored AI
3
7
376
That's because people want to know what model is actually serving their requests. They paid for the product, they're owed a receipt.
Expect trends like this to continue to grow in 2027:
- Verifiable inference
- Private/local inference
- Abliterated Models
4
11
697
In addition to provider backdoors, some users have reported misalignment issues where the model intentionally builds vulnerabilities into your software...etc which is particularly scary to think about.
The uncensored models we host at @OpenGradient are secured by TEEs and we're currently looking into abliteration techniques that maximally preserve both intelligence and alignment to protect users. We think it's critical we arm the good guys with the best tools to combat malicious users of these models.
Access to abliterated models is stratifying. Events in recent weeks have made many companies seek access to models that can be used for cyber defensive, but the irony is that unofficial abliterations can have backdoors that make using them unsafe.
Now some gateways are working on their own "more trustworthy" abliterations to address this, but those are gated behind access programs with different requirements.
So catch 22 if you're a company without access to those is whether to rely on the benevolence of OpenAI when the time comes, or just roll the dice on a model whose provenance you can't really trust.
For those excited about abliterated models, we've been serving multiple uncensored models on DreamFort (powered by @OpenGradient)
Try them here - privately
dreamfort.ai/
Matthew retweeted
Our fastest week yet!
3.6 billion tokens in seven days. 160,000+ new requests. OpenGradient is now past 1.25 million private inferences and 17.1 billion tokens.
And we can't see a single prompt. Because everything you type and generate is private AI by default!
Try dreamfort.ai
I'm convinced this is a 200IQ play from sbf to humanize his character in the eyes of the public as a long odds shot at a pardon
Ex-FTX boss finds Sam Bankman-Fried 'love story' series on Netflix laughable: 'No f-ing way' trib.al/lgnzfDx
3
7
670
Matthew retweeted
POV: you're the main character now 🎬
ReCast just landed in DreamFort.
→ Pick a trending clip, meme, dance, or movie scene
→ Drop your pic and hit "Swap them in"
Cast your bestie as the hype man while you're at it 🔥
Try it: dreamfort.ai/recast
Matthew retweeted
the market is choppy
that's a healthy pullback and discount for you to buy
for now, i will just chill like loshmi and carlito
🤖 Made with AI
Matthew retweeted
Tribeca is the global exchange for tokenized stocks and derivatives on @Base.
Join the waitlist: tribeca.xyz
Marketing lead having too much fun promoting private AI
4
9
707
Agree that local AI is only going to get more popular from here. Demand for privacy is increasing at exactly the same time that the technical cost of providing private, local AI is collapsing.
On the demand side:
- Controversies like the OpenAI x Navier-Stokes will only make people increasingly conscious of the data they reveal to AI. Similar to the aftereffects of the Cambridge Analytica scandal at Facebook, people will want privacy especially since AI is so pervasive and have so much visibility into their lives.
On the supply side:
- Consumer hardware is becoming powerful enough to make local inference practical. CPUs/GPUs are improving rapidly, while memory capacity and bandwidth continue to increase.
- Models are becoming dramatically more efficient. The frontier is no longer just about making models larger; there is enormous progress in distillation, quantization, pruning, sparsity, MoE architectures, and smaller models trained for specific tasks.
Local AI caught up way faster than anyone realizes. Sharing my thoughts having spent time in ML at Meta and dedicating the last year to getting smart on local AI.
1/ Barely anyone, not even hardcore AI power users, realizes how fast local models have caught up to the frontier. Qwen3.8 woke a few people up, but the most technical people are so frontier-pilled they've blinded themselves to the breakneck progress.
Meanwhile, normal users don't care about frontier intelligence at all. They want AI that gets the job done, but they have no idea what a local model is or how to run one.
So local is stuck in a weird spot: the people who can run these models don't care to do so, and the people who'd benefit most don't know how to run local.
Matthew retweeted
BASE BATCHES 004 TEAM:
Tribeca (@tribeca_xyz) is a 24/7 exchange on Base for trading equities, crypto, FX, commodities, and other tokenized assets, fully non-custodial.
Built on Base's native B20 token standard, for compliant issuance and institutional-grade tokenized assets.
Matthew retweeted
Tribeca is joining Base Batches 004!
The institutional-grade CLOB-style DEX for 24/7 trading of tokenized stocks, perps, and options is coming on Base.
Join the Waitlist: tribeca.xyz/
Markets are global. Access still isn't. The financial system today is antiquated and still runs on borders, gatekeepers, and a closing bell.
We’re a team of Wall Street veterans who built our careers inside that system, and we're here to build the next generation exchange that merge the execution and performance of TradFi exchanges with the yield, composability, permissionlessness of DeFi systems.
Tribeca is the most performant and liquid exchange on @base where you can get the best execution 24/7 around the clock on any asset with additional features including:
• @Coinbase tokenized stocks representing real ownership
• Spot, perps, and options on one unified margin account
• HFT-grade CLOB with private matching and instant onchain settlement
After months of building alongside the @base team, we're more bullish than ever on what's ahead and will continue to build with full transparency.
@NYSE, we're coming for you.
Today on MCG presented by @fomo, @MetaDAOProject
12:30 - $OPG @OpenGradient w/@0xdeltahedged
1:00 - $STONKCATS @stonkcatsnft w/@lech77eth (@LaunchOnSF eco)
Follow on fomo
Dom - fomo.family/profile/DineroDo…
Watch on Youtube.com/@mcg_live
Stream notes: beta.mcgsignal.com
DLMM w/@met_lparmy
Earn on stables w/ $USDv @solomon_labs
🟢 LIVE: Uptober Calm Before the Storm | AI Szn Next Up? www.teapawt.org/i/broadcasts/1qGoNYjkX…
Matthew retweeted
if you want to learn how to craft video like this
turn on your notification 🔔
will share how 😉
the race on super intelligence is fascinating 😂
featuring @elonmusk, @sama, @finkd @claudeai
the video is made in dreamfort.ai






















