@heymikasagi @powersync_ @supabase @tan_stack @k081e @devagrawal09 @barnesmichal here to enable great AX 🫡
English
2027.dev
32 posts
2027.dev
@2027dev
Make something agents want
San Francisco Tham gia Şubat 2026
4 Đang theo dõi47 Người theo dõi
What great Agent Experience really unlocks: new production-ready apps built in one shot using @powersync_, @supabase and @tan_stack
Honored to be part of this journey with @2027dev 🔥
Congrats on all the progress @k081e, @devagrawal09, @barnesmichal!
Kobie sync/acc@k081e
Just over a year ago we started testing vibe coding tools like bolt.new and our AX was really really bad. I'm talking hallucinations even trying to install the PowerSync JS SDK. Fast forward to today and you can now zero-shot a new app that uses @powersync_ This is the result of a lot of hard work across our entire team. Watch @devagrawal09 show us the state of things. p.s. we're still investing in AX, now turboboosted by the folks at @2027dev 🥂
English
frictionless AX is the ultimate goal 🔥 congrats, @powersync_ and @k081e!
Kobie sync/acc@k081e
Just over a year ago we started testing vibe coding tools like bolt.new and our AX was really really bad. I'm talking hallucinations even trying to install the PowerSync JS SDK. Fast forward to today and you can now zero-shot a new app that uses @powersync_ This is the result of a lot of hard work across our entire team. Watch @devagrawal09 show us the state of things. p.s. we're still investing in AX, now turboboosted by the folks at @2027dev 🥂
English
Just over a year ago we started testing vibe coding tools like bolt.new and our AX was really really bad. I'm talking hallucinations even trying to install the PowerSync JS SDK.
Fast forward to today and you can now zero-shot a new app that uses @powersync_
This is the result of a lot of hard work across our entire team. Watch @devagrawal09 show us the state of things.
p.s. we're still investing in AX, now turboboosted by the folks at @2027dev 🥂
PowerSync@powersync_
Watch Codex zero-shot Supabase and PowerSync integrations into a TanStack Start app using nothing but the CLIs!
English
100+ saves in ~24h!
Glad to know we wrote something people wanted
more soon @2027dev
Mika Sagindyk@heymikasagi
English
4 weeks ago, we launched Agent Arena on @2027dev
Since then, we ran Agent experience (AX) evals on 50+ devtools
One single task: evaluate how easily AI agents can set up tools, fully autonomously
Here are five things that surprised us 🧵
English
English
Biggest learning from this release: there are way more LLM observability platforms that we realized!
stay tuned for an updated ranking: adding @cometml, @tryadaline, @agenta_ai, @lmnrai & @wandb soon
comment/DM if I'm forgetting anything!
+ drop another category suggestion anytime :)
Mika Sagindyk@heymikasagi
New evals on Agent Arena: ✨LLM observability✨ AX (agent experience) ranking as of 3/03/36: 1/ @raindrop_ai 2/ @langfuse 3/ @langchain 4/ @RespanAI 5-6/ @PortkeyAI, @braintrust 7/ @helicone_ai (now part of @mintlify) 8/ @arizeai More on 2027.dev/arena We measure how easy it is for AI agents to get started with devtools, fully autonomously With AI agents becoming the primary consumers of docs and APIs, AX is the natural evolution of DX DM me for your full AX eval! If you're missing a tool or category, comment below cc @benhylak, @alexisgauba, @snarkyzk, @marcklingen, @maxdeichmann, @nimarblu, @hwchase17, @ankush_gola11, @samecrowder, @Andydy42, @raymond_huang26, @jumbld, @ankrgyl, @daRubberDuckiee, @justinstorre, @coleywoleyyy, @jason_lopatecki, @seldo
English
make something agents want
Mika Sagindyk@heymikasagi
New evals on Agent Arena: ✨LLM observability✨ AX (agent experience) ranking as of 3/03/36: 1/ @raindrop_ai 2/ @langfuse 3/ @langchain 4/ @RespanAI 5-6/ @PortkeyAI, @braintrust 7/ @helicone_ai (now part of @mintlify) 8/ @arizeai More on 2027.dev/arena We measure how easy it is for AI agents to get started with devtools, fully autonomously With AI agents becoming the primary consumers of docs and APIs, AX is the natural evolution of DX DM me for your full AX eval! If you're missing a tool or category, comment below cc @benhylak, @alexisgauba, @snarkyzk, @marcklingen, @maxdeichmann, @nimarblu, @hwchase17, @ankush_gola11, @samecrowder, @Andydy42, @raymond_huang26, @jumbld, @ankrgyl, @daRubberDuckiee, @justinstorre, @coleywoleyyy, @jason_lopatecki, @seldo
English
we're back with new AX evals! 2027.dev/arena
Mika Sagindyk@heymikasagi
New evals on Agent Arena: ✨LLM observability✨ AX (agent experience) ranking as of 3/03/36: 1/ @raindrop_ai 2/ @langfuse 3/ @langchain 4/ @RespanAI 5-6/ @PortkeyAI, @braintrust 7/ @helicone_ai (now part of @mintlify) 8/ @arizeai More on 2027.dev/arena We measure how easy it is for AI agents to get started with devtools, fully autonomously With AI agents becoming the primary consumers of docs and APIs, AX is the natural evolution of DX DM me for your full AX eval! If you're missing a tool or category, comment below cc @benhylak, @alexisgauba, @snarkyzk, @marcklingen, @maxdeichmann, @nimarblu, @hwchase17, @ankush_gola11, @samecrowder, @Andydy42, @raymond_huang26, @jumbld, @ankrgyl, @daRubberDuckiee, @justinstorre, @coleywoleyyy, @jason_lopatecki, @seldo
English
@heymikasagi @raindrop_ai @langfuse @LangChain @RespanAI @PortkeyAI @braintrust @helicone_ai @mintlify @arizeai make something agents want
English
New evals on Agent Arena: ✨LLM observability✨
AX (agent experience) ranking as of 3/03/36:
1/ @raindrop_ai
2/ @langfuse
3/ @langchain
4/ @RespanAI
5-6/ @PortkeyAI, @braintrust
7/ @helicone_ai (now part of @mintlify)
8/ @arizeai
More on 2027.dev/arena
We measure how easy it is for AI agents to get started with devtools, fully autonomously
With AI agents becoming the primary consumers of docs and APIs, AX is the natural evolution of DX
DM me for your full AX eval!
If you're missing a tool or category, comment below
cc @benhylak, @alexisgauba, @snarkyzk, @marcklingen, @maxdeichmann, @nimarblu, @hwchase17, @ankush_gola11, @samecrowder, @Andydy42, @raymond_huang26, @jumbld, @ankrgyl, @daRubberDuckiee, @justinstorre, @coleywoleyyy, @jason_lopatecki, @seldo
English
2027.dev đã retweet
Hey sf! we're doing smth fun -- hosting a meetup for AI agents this wednesday🦞
Bring your claude code (opencode/codex/openclaw also allowed)
Format: 5min lightning talks: show us your vibe coding setup (agents, workflows, what's actually working)
come hang! DM if you want to present your flow
up to ~35 people, 5 lightning talks
Food & drinks included
Hosted by @checklyHQ × @2027dev; cc @HLENKE
English
2027.dev đã retweet
@jeff_weinstein these companies are doing solid agent-first work: @tavilyai, @e2b, @supabase, @clerk, @togethercompute
(top in their categories)
worth following the builders behind these products!
we've been evalauting what makes devtools agent-friendly on 2027.dev/arena
English
Agent Arena update: new Sandbox evals just dropped!
added new providers: @vercel , @modal and @Cloudflare
New AX (agent experience) ranking as of 02/25:
Grade B: @e2b, @vercel, @daytonaio & freestyle.sh
Grade C: @Cloudflare, @blaxelAI, @modal & @codesandbox
More on 2027.dev/arena✨
We measure how easy it is for AI agents to get started with devtools, fully autonomously. DM for your AX report 🫡
Shoutout to @benswerd who took action from our AX report and improving the freestyle score in <24 hrs
cc: @mlejva, @tereza_tizkova, @ivanburazin, @jaysaadana, @rauchg, @andrewqu, @craigsdennis, @paul_s3i, @charles_irl, @bazzjuh
English
@heymikasagi @vercel @modal @Cloudflare @e2b @daytonaio @blaxelAI @codesandbox here to deliver daily AX evals 🫡
English
“It’s 2026. Build for Agents!”
So true and exactly what we’re seeing developers wish for: tools that are designed for agents, not just humans
The caveat is that it’s not as easy as it may sound
When we design systems, we tend to build for human intuition rather than machine logic. Simply keeping agents "in mind" isn't enough; it requires rigorous testing and evals to ensure coding agents can actually navigate our systems
AX matters more every day. Here to deliver live ranking of how agent-ready devtools are on 2027.dev/arena 🫡
Andrej Karpathy@karpathy
CLIs are super exciting precisely because they are a "legacy" technology, which means AI agents can natively and easily use them, combine them, interact with them via the entire terminal toolkit. E.g ask your Claude/Codex agent to install this new Polymarket CLI and ask for any arbitrary dashboards or interfaces or logic. The agents will build it for you. Install the Github CLI too and you can ask them to navigate the repo, see issues, PRs, discussions, even the code itself. Example: Claude built this terminal dashboard in ~3 minutes, of the highest volume polymarkets and the 24hr change. Or you can make it a web app or whatever you want. Even more powerful when you use it as a module of bigger pipelines. If you have any kind of product or service think: can agents access and use them? - are your legacy docs (for humans) at least exportable in markdown? - have you written Skills for your product? - can your product/service be usable via CLI? Or MCP? - ... It's 2026. Build. For. Agents.
English