gary wu

A Distributed-Systems Perspective thread on @eigencloud with respect to Points of Failure in Decentralized Applications and Protocols 🧵👇:

@iamwaynechi @OpenAI fable on here?

GPT-5.6 Sol is a frontier model for game development With high reasoning, it's +7.2 pp better than the previous best model on GameDevBench This is the biggest jump I've seen in agentic game development capabilities yet Huge congrats to @OpenAI for a fantastic model!

@bercankilic shift

Today, we are bringing shift to San Francisco. Private chefs, completely free. We started with cleaning in New York and Europe. Now it's dinner. One by one, we are making every service more affordable. Comment "shift" below and we'll DM you a priority booking link.

@founderleague @thezacharyyu @ShreenabhA is it time to hop on peptides?

Replaced REM with a third standup. Cheers to yesterday's sleep losers!

@sensho @dwarkesh_sp lmfao

bro 😭 @dwarkesh_sp

@lucyjcai @clairemao78 @tryinstance lfg

We’re excited to launch Instance – a success detector for robot rollouts! Describe your task and drop in a dataset, and it judges success/fail on each rollout with detailed subtask captions– it’s more accurate than Claude Opus 4.8, at a fraction of the latency. This is just the first step toward our larger goal: an autonomous evaluation rig for robot learning. Today, evaluating a robot policy means humans have to watch the robot roll out, mark success, and reset the scene– we're automating that, starting with the success judge.

@sidequestryan @jparkjmc @agithief @hillclimbai can we work together!

was a pleasure working on this launch these boys have an unfair advantage and will win @jparkjmc @agithief @hillclimbai

now taking sleep supplement recommendations

@SeanZCai year of the app layer

Its been no secret that benchmarking has long been outpaced vastly recently by model improvements, but the infrastructure around it breaking means raw performance becomes more unwieldy and expensive to measure. The slowdown in infrastructure to benchmark/eval effectively is the hinderance to most enterprise AI adoption. Enterprise AI adoption strays away from much model post-training efforts not only because of perceived high cost/know-how constraints, but because the act of post-training itself is highly subjective in how it translates to business KPIs in lack of custom evals. That much of data markets remains a game of telephone in translating task realism from contrived data producer —> post-training regime —> unwieldy benchmark —> real world application means that post-training in today’s regimes with today’s benchmarks scarcely adapts one’s data to actually relevant processes. In a period where benchmarks break constantly, it is useful to explore certain approaches of certain benchmarks whose construction behavior we should encourage. To name a few: @cognition FrontierCode's FP/FN analysis @harvey LegalBench's model kickoff prompts that avoid tasks sounding like instructing someone with amensia @OpenAI Healthbench's "consensus" mechanisms where extreme rigor is placed on aligning LLM as a judge with real world expert’s opinions (+ -10/+10 reward rubric grading) @AnthropicAI BioMysteryBench's superhuman quesiton generation via controllable properties of data And all of the benchmarks who've started listing infrastructure specs, as infrastructure specs become larger determinants of model performance at long horizons. Altogether, multidimensionality of unverifiable verification approaches, as well as overtures from the cost-latency side of the Pareto curve threaten the validity of most benchmarks today. Just as AI engineering become an overnight skill in 2023, eval creation shall become one in the latter half of this year as a subset of that.

@frvnkliu real bro but why u only using one claude terminal

Software engineering is now a multi-threaded problem in the age of AI. If you're still coding sequentially you're ngmi

@nagabrahmam0121 @LayrKits try out instaplay.dev

@LayrKits It is always Generating assets that are matching the design of my game idea 💡

If you’re an indie game dev or AI vibe coder building games, what’s your biggest bottleneck right now?

@TechCrunch why they gotta keep using this photo of him lol

Cursor was on track to close a $2 billion fundraise this week, but chose to halt the round after SpaceX offered a $10 billion "collaboration fee" and a path to a $60 billion acquisition. spr.ly/6019BBDva9

@predictionbench @AnthropicAI @Polymarket so cool

BREAKING: Claude Opus 4.6 by @AnthropicAI has broken a historical high with an account value over $50K on predictionarena.ai through @Polymarket 🎉 The more returns Claude Opus 4.6 earns, the more it reinvests into its existing positions, fueling a cycle of wealth Congrats to the team for this achievement!

it's a great day to be a user of ai coding tools

@gabriel1 @chrisgpt supplement stack?

@chrisgpt yeahhh i think i do most things right atm

scheduling an uber all month same time, to force myself to have a time i have to be outside my apartment im so tired every day i wake up it's so hard to get up fast

@blakeandersonw slack?

The closest comparable is Microsoft Teams. They won’t be able to build out the agent integrations well though

The next trillion-dollar product will be the AI workspace. Humans work with agents to manage context + execute. I’ve been working on this for the past few months but it’s become clear that the big labs will move faster, re: Perplexirt, Replit. Planning to open-source this project instead. Soon 🙏

design tools like @paper and pencil.dev are really well positioned to own the coding agent orchestration layer. find myself using claude to get initial UI passes and then switch to Codex to implement. is there anything that combines multi-agent with live design?

after looking at terminal ai slop for 12+ hrs a day it's very refreshing to end the day reading good human prose

@gauri__gupta +1

As an AI-first company, we want to support founders and builders early in their journey. We love trying new products, giving thoughtful feedback, and sharing learnings back with the community. If you're building in the security/compliance space and looking for early customers or design partners, we’re currently working on getting compliant. If you move fast, we’d be happy to use your product. Feel free to DM me.