Ryan Fowler @ryanf_coder

everyone else stores. we learn @metamemoryai is memory for AI agents that gets smarter the more you use it🧠

It's hard to fathom that in ~12 months, we'll likely have access to a model 2x more powerful than Fable 5. And that might be the conservative estimate, & model costs will likely be way cheaper at that point, too. Here's why: The best single measure of AI capability is Epoch's Capabilities Index (ECI) (a composite score across 40+ benchmarks - math, coding, science, reasoning). Think of it like an IQ test for frontier models. When Fable was released, it hit a new high of 161 on the ECI. Now here's the part that matters: The rate of frontier improvement nearly doubled around April 2024, from about 8 points per year to 15. That faster pace has now held across every single quarterly measurement cycle since then, and is the new trend. So, in theory, the trend that's held since 2024 just needs to continue in post-Fable 5 model releases, and the floor in 12 months is a 2-3x improvement in model capability. But the wildest stat isn't even the capability curve. It's the cost curve: The cost to run an AI model at a fixed level of performance is halving every 2 months. Put it together: In 12 months, you won't just have a model 2x more capable than Fable 5. You'll have it cheaper, faster, and possibly running on hardware you already own. We're so early.

Something quietly inverted in AI compute this year, and it changes what the buildout is actually for. In 2023, 2/3 of AI compute went to training, the actual work of building a model. The other, smaller slice went to inference, the work of actually running it once it's built. But that ratio quietly started flipping. Inference is now 2/3 and still climbing, per Deloitte, and the chips built to run it crossed $50B this year. The main reason this flip matters (and it's not percentage-wise): training and inference are different animals. Training happens in bursts, on one giant cluster, then it's done. Inference never stops. It runs every time someone sends a prompt or an agent takes a step, and it scales with every user you add. One is a construction project. The other is a utility bill that grows forever. Every assumption about AI infrastructure was built around training, because that's where the money went. The money just moved to the workload that doesn't need to sit in a single cluster to run.

259,783 packs opened in one week on @Collector_Crypt Roughly one every 2.3 seconds This is what it looks like when Solana turns multiple economies into one smooth consumer flow.

Sailor @solana_sailor

11 hours ago

The best Solana apps are starting to sit between categories A pack opening flow is a clean example Frontier showed the same direction with winners like @Rosentica bringing TCG closer to gaming, consumer and ownership Wallets are moving there too Solflare Packs, powered by

17 12 63 5K 10

We are going to switch up some things and ship a lot more on growth. Stay tuned!

Umbra Privacy @UmbraPrivacy

10 hours ago

Every pool on Umbra has a Privacy Score that reflects its overall anonymity. The more users and transactions a pool has, the stronger the privacy it provides. Know your privacy before you transact.

8 5 55 5K 3

You can now access 224+ free AI APIs from Google, Groq, NVIDIA, OpenRouter, Mistral, Cohere, and more. freeLLM is a directory of free AI models from 25 providers, complete with rate limits, context windows, API key links, and one-click config snippets. Works with Claude Code, Cursor, Codex, OpenCode, Hermes, and pretty much any OpenAI-compatible client. Setup: – Grab API key: freellm.net – Select a model (e.g. cf/moonshotai/kimi-k2.7-code) – Generate a free API key – Plug it into OpenCode, Hermes, Cursor, etc. If you're building agents, coding, or experimenting, this is basically free inference. Lock in and start building today. Thank me later.

one of my friends tell me that the @tokens skills installs only to @claudeai folder .claude ... why not .agents for Codex.. and other models... i said to him hold my zobo....🤣🤣😂😂😂

Genuinely striking data from OpenAI on how fast agentic work is taking over. Non-developers adopted Codex faster than developers did, up 137x for individuals in under a year. Legal, Finance, and Recruiting all flipped to it as their primary tool. The thing that makes it work isn't coding ability, it's that people can finally direct technical execution in their own domain without the translation layer in between. The barrier to doing technical work was never the domain expertise. It was everything between the expert and the machine and thats collapsing fast.

OpenAI @OpenAI

2 days ago

Work at OpenAI is being transformed by agents, in every department. Across our entire company, people are using Codex to do work that is more complex, longer-running, and increasingly cross-functional. Our internal usage offers an early look at how agentic tools may reshape

keyboard_arrow_left Previous keyboard_arrow_right Next

400 443 5K 1.2M 2K

The future of AI is undoubtedly open-sourced. For 99% of people, the top open-weight models can already handle all your daily needs. Imo, the best thing you can do right now is run a "barbell" approach for your AI usage: Intial 10% and planning phase: Use frontier intelligence (Opus/GPT) Gruntwork and the middle 80%: Switch to open-weight model (GLM, Kimi) Final 10% and review: Switch back to froniter intelligence to verify Best of both worlds - cost efficiency + intelligence.

best chart in crypto, 6 years running.

Colosseum @colosseum

a day ago

1/ Announcing the winners of the Solana Frontier Hackathon!🏔️ Read about the winners & honorable mentions: blog.colosseum.com/announcing-the… The subset of winning teams accepted into our VC fund's next accelerator cohort will be shared in the coming days. Congrats to all! 🏆

311 171 1K 463K 185

Claude limit hit every 5hrs. Lost work multiple times. Fought bugs. Just shipped Velocity AI: an autonomous lead acceleration engine that calls inbound leads in seconds, qualifies them + updates your CRM while your team sleeps. 🔗omnivoiceai.vercel.app Watch the demo below 👇

Matt Sim📍Network School @RVAClassic

3 days ago

day 116 of my @ns journey >worked all morning + day on submitting hackathon for @omnidimAI (failed + restarted lots) >spent my last dollars on @screenstudio >secured @SuperteamMY grant for @OperatorUplift >looking for clients for launch videos >claude pro plan sucks >laundry day

2 0 12 9K 2

I made a football edit with @SuperteamNG leads and members Whose animated PFP can you spot? 🤗 Animated with @ofiboxstudios

Superteam Nigeria @SuperteamNG

5 days ago

The Solana Ecosystem Call Anime World Cup is here and we’ve joined the competition ⚽️🏆 Choose our team on the @SolEcoCall Footify app and the real Superteam Nigeria jersey is yours 🇳🇬 Generate your jersey here: Footify.us Sign up on Luma: luma.com/0j8gx7i1

keyboard_arrow_left Previous keyboard_arrow_right Next

55 37 218 49K 26

It's trending 😯📈

Base Build @buildonbase

5 days ago

Builder Codes are live for x402 on Base. With Builder Codes, teams building with x402 can: → Unlock app-level analytics → Attribute x402 traffic to your project → Integrate in minutes with a coding agent Here’s how to get started ↓

76 100 449 189K 131

i gave my agent one of my @tokens API KEY guess what. he got the skills and he actually use it. go and get the @tokens skill for your agent here. npmjs.com/package/tokens… npx tokens-xyz-skill with this magical word you can add it to your project.

Iraq to OPEC: increase our oil production quota — or else!!! bloomberg.com/news/articles/…

Become your team's MVP: footify.us Sign up on Luma: luma.com/0j8gx7i1?tk=Zb…

1,100+ Player Cards and counting. We'll be awarding a Breakpoint ticket to each of the top 3 teams. You have until tomorrow's Solana Ecosystem Call to climb the leaderboard. The clock is ticking. 👀

Get a wallet's balance at any point in time The new balance-at endpoint returns a wallet's exact SOL or SPL token balance at a given date or slot How much USDC did a wallet hold at 19:20 on Jan 10, 2025? Now it's a single call Ideal for PnL, charts, snapshots, and cost basis

Historical token balances are now live on Orb. View how a wallet's balance and value for any token has changed over time. Enjoy!

2026 World Cup prediction markets have crossed $5B in trading volume @Polymarket and @Kalshi together recorded over $5B across match bets, tournament winners and player markets For comparison, Polymarket recorded only around $138k during the 2022 World Cup Prediction markets have gone from a niche product to a multi-billion-dollar sports market