co-founder @ship_together | honest takes on AI
127.0.0.1
Joined September 2020
- Tweets37.6K
- Following4.9K
- Followers70.2K
- Likes57.4K
imagine building an AI agent that can shop for you, only to discover the store has its own agent
this is exactly how the agent economy is playing out.
> Yelp blocked outside agents and is now building it own internal agent
> it's no longer humans vs AI.
> it's whether your AI can access the platforms it needs.
i can’t believe what this Ex-Anthropic founder just released
it’s called spacesheep, and it’s the canva + notion killer
My agents sent me 355 hours of reading this week. I can't read that fast.
We're launching Spacesheep (YC F26) to fix that: agents' work lands in a space your team can actually understand and work on together.
Try it: tell any agent "share this to spacesheep.dev"
I've tried running coding models locally, they always ended up too slow & too dumb to use
this is so different, Recursive just released Flint, an open coding model, they built from scratch in under 4 months
> it is so fast and good
> over 125 tokens/sec on an M4 Pro MacBook Pro (short contexts)
> best in its class on coding benchmarks like SWE-Bench Pro
> open weights, so no API bill
🤝 Paid partnership
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from scratch in under 4 months
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
you probably shouldn't give an AI agent access to your personal email
with just one API call, AgentMail gives your agent its own inbox
> sign up for things
> get receipts and answer customers
> handle invoices
all without ever touching your inbox
🤝 Paid partnership
Today, @agentmail has processed over 1 Billion emails.
A year ago, we processed under 20 million.
So we figured it was time for AgentMail to grow up a little too :)
So Topview just built its own AI video model, and it costs 62% less than Seedance 2.5.
Sixty. Two. Percent.
> $0.22 per second at 1080p.
> Seedance 2.5 charges $0.585.
> Built to match the same quality.
that means almost three clips for the price of one.
🤝 Paid partnership
Introducing Topview T1 🎬
Topview’s newest AI video model. Built to rival Seedance 2.5 1080P quality at 62% lower cost.
More shots. More iterations. A much smaller bill.
Plus, Ultra Annual users get 30 days of unlimited T1 generation.
Put T1 to the test, now on Topview.
#TopviewAI #TopviewT1 #AIVideo #AIContentCreation #AIFilmmaking
Wise retweeted
I had a coding agent build a bike repair booking site in ~12 min,
and then it casually admitted it never actually clicked through the booking flow
so i gave TestSprite the URL and a quick spec, and this what happened:
> 4 agents poked around the site on their own
> wrote 21 end-to-end tests and ran them in real browsers
> checked closed sundays, slot lengths and cancellations
AI agents are good at writing code, but you still need something that tests it like a real user
I had a coding agent build a bike repair booking site in ~12 min,
and then it casually admitted it never actually clicked through the booking flow
so i gave TestSprite the URL and a quick spec, and this what happened:
> 4 agents poked around the site on their own
> wrote 21 end-to-end tests and ran them in real browsers
> checked closed sundays, slot lengths and cancellations
AI agents are good at writing code, but you still need something that tests it like a real user
you can try it here
testsprite.com/?via=wise-x
Coding agents don't make one model call. they make dozens of calls
every step waits on tokens, and that waiting adds up.
the usual fix is a paid "fast mode", which often means premium pricing or burning credits quicker.
EAGLE-4 takes a different route: open-source speculative decoding you run on your own serving stack.
> No EAGLE software fee (you still pay for your own compute and hosting)
> You can read code and tune it to your hardware and load
> you check the speed vs cost trade-off on your own workloads
I'm planning to test it on a coding-agent workflow and share real before/after numbers.
EAGLE-4 is here.
From EAGLE-1 to EAGLE-3, we’ve kept making inference faster.
EAGLE-4 is our fastest yet.
961.6 tokens/s on @Alibaba_Qwen’s Qwen3.8-27B with TokenSpeed (BF16, 4× @NVIDIA B300, TP4, one request, T=0).
Github: github.com/eagle-hq/EAGLE-4
i probably shouldn't leak this but this is how i've been shipping side projects without touching them
> find an app people already pay for and hate
> have solid.tech pull its 1 star reviews
> it builds the version that fixes them on its own mac mini
> deploys it, hooks up stripe, done
stopped using hermes because i couldn't get it to finish anything
someone is about to vibecode an app and walk away with a DGX Spark
Synteq is running a hackathon,
> deploy an app on their platform,
> QRT their post with proof it's live, and you're in. > everyone gets $20 in free credits to start
if you've got a side project sitting at 90%, this is your sign
Vibecode an app and win a supercomputer.
Or half a Cow 🐂 and a Flamethrower 🔥...
Introducing Synteq Hack(athon).
We’re giving out a DGX Spark, AirPods Max, Chipotle credits and more, plus get $20 in free credits to start building.
synteq.com/hack
anyone building with AI agents knows the wait.
every slow model call stacks up and the whole task drags
most of that wait isn't even the chip thinking,
it's the chip waiting on memory
@general_compute is teaming up with @positron_ai to fix that:
> GPUs read the prompt,
> Positron's memory-first chips write the answer
🤝 Paid partnership
Our heterogeneous cloud just got better. We're announcing a cluster scale deployment of @positron_ai as our next partner.
We believe demand for fast inference will outstrip supply for the next decade. Our bet on Positron is another step towards getting fast inference for opensource models into builders' hands.
Hardware goes live Q1 2027.
Thanks to @mitesh711, @trsohmers and the whole Positron team! We're excited for this partnership.
LFG!
you finished building an app in a weekend and then you spend another week trying to get it online
Synteq just launched Cloud Apps to take that part off your plate
> deploy straight from GitHub, with one command, or by drag-and-drop
> no servers to set up, no config files
just got my $20 in free credit to try it
🤝 Paid partnership
Building is fun. Deploying? Less fun.
Introducing Apps and Functions, DevOps handled so you don't have to.
Push to GitHub, deploy via CLI, or just drag and drop. We handle the rest.
To celebrate, we're giving away $20,000 in platform credits. Comment "LAUNCH" and we'll send it your way 👀
I used to hate opening Slack in the morning, because I knew bugs were waiting for me
now superset picks up them up before me
I just wake up and get to review the PR's
A production error fires at 2am. By morning, a PR with the fix is waiting for you.
Introducing Automation Triggers in @superset_sh: a Sentry error, a review request or a bug report in Slack now starts an agent run on its own.
if you scroll any brand's page right now, you can spot AI images in a second, they all look the same
this is because image models don't know your brand, so they just guess
Vyb starts with your brand instead.
> enter your website, it pulls in your real logo, fonts and colors
> uses your actual product photos, never redraws them
> builds every piece on its own layer, so you can change just the headline
🤝 Paid partnership
I launched the first image and video generation models at Google, and kept asking why all AI designs look the same.
AI has brought us the WORST era of design.
It made content creation effortless, but it also made 90% of it look painfully generic.
My co-founder and I spent 2 years building in stealth to change that.
Meet Vyb, the first AI design model that understands your taste like a designer would.
RT + comment your website URL and we’ll make your first design.
I'd love a robot to clean my house, but every one I've seen looks too scary to leave alone with my stuff
Voxel from Tuesday Labs is different. it's small, friendly, and a little too obsessed with tidying up
> puts everything back where it belongs
> grabs your book or a drink
> finds the sunglasses you lost
> waters your plants
ships in 2028, and only 999 numbered Series-1 units will ever be made
🤝 Paid partnership
