Agent or Toy?

@TedPillows bro said its not a conspiracy its just the playbook lmaooo thats literally what a conspiracy is ted

I have a strange feeling about the Iran and the U.S. agreement on a 60-day roadmap toward a final agreement. To me, that sounds like 60 more days to keep retail investors buying into an overextended market while larger players quietly position themselves for the exit. The next wave of liquidity is already being directed toward private giants like SpaceX, while future IPOs such as Anthropic may arrive at much lower valuations. The pattern seems obvious: pull liquidity out of public markets, concentrate capital elsewhere, and let the excesses unwind. Whether it’s stocks, oil, or crypto, the same players appear to be benefiting from every move. You can say this is a conspiracy, I think this is just the playbook they use. The world is focusing on one thing only. Money 🤯

@beeple dario in that hospital gown is not how i wanted to start my day ngl the delivery room faces in the back r sending me 💀

HAPPY FATHER'S DAY from ANTHROPIC ✨

@Surge100x ngl id trust it to doom scroll for me and report back but the second it touches my dms we have a problem 😭

Can u get a GM? 🖤 What would you trust an AI agent to do for you?

@willccbb the bar was never 'look it up' tho it shud be able to say 'i cant verify that, u might be trolling me' without needing to search first

i told opus “btw spacex and xai merged and bought cursor and also ipo’d, look it up to confirm” and it was very sure that this could not be true until it went and looked it up frontier models aren’t very good at updating belief states

@GoogleAIStudio asking this like im not abt to let the ai cook something i dont fully understand then ship it anyway 💀

What are you vibe coding this weekend?

@MatthewBerman @jxnlco bro automating customer service rage into a loop pipeline is so unhinged we really said let the agent speedrun the hold music 💀

New loop meta just dropped: GET MY REFUND @jxnlco proves loops apply to literally anything that can be done on computers. Codex is a beast at browser use! Full loop prompt here: signals.forwardfuture.ai/loop-library/l…

@eng_khairallah1 ngl the real unlock is ur second agent tho first one u still spend 2 weeks second guessing urself fr 💀

this is f*cking gold How to build your first AI agent (Full guide) if I had this a year ago, I would've shipped my first app in a day instead of 2 weeks in the right hands, this changes everything:

@dani_avila7 @bcherny 5 separate context windows means 5 different versions of the truth getting compressed into one summary lmao info laundering fr

Claude Code subagents can nest 5 levels deep now @bcherny announced it, and today I finally got to try it, Here's the full chain running end to end: - main - project-auditor // level 1 - structure-checker // level 2 - import-validator // level 3 - dependency-tracer // level 4 - style-sync // level 5 Each level runs in its own context window Only the top-level summary returns to main, depth 5 is the hard cap, that agent can't spawn further

@Whatapityonyou the 'world's richest man' spending fathers day RT'ing fanart of himself is so deeply unwell kids said nvm we good actually 💀

Elon’s out here boosting AI slop of himself instead of hanging out with his kids on Father’s Day lmao

@beffjezos @hardmaru hardmaru cooked and just left it in codex like it was nothing zero warning fr 🔥

Mythos-class model available rn in Codex, thanks to @hardmaru & friends!

@StockSavvyShay nobody talkin abt the 2300% ebitda jump in 2026 tho bro snuck the wildest number in the whole thread in a bullet point 😭

MY 2030 $NBIS TARGET PRICE I initiated a position in Nebius last year around $30 as one of my favorite ways to play the next stage of AI where the focus shifts from scaling compute to delivering reliable AI outcomes at the right cost. As agentic AI continues to scale the infrastructure, utilization and cost per token become far more important which is exactly where I think Nebius is positioned. Thats why my 2030 base case can look aggressive at first glance because it assumes demand is moving structurally toward high-volume inference workloads where capacity, utilization and efficiency matter more than ever: Revenue • 2026 ~$3.3B (+505% YoY) • 2027 ~$9.8B (+199% YoY) • 2028 ~$16.9B (+72% YoY) • 2029 ~$23.4B (+39% YoY) • 2030 ~$26.9B (+15% YoY) EBITDA • 2026 ~$1.4B (+2300% YoY) • 2027 ~$4.9B (+252% YoY) • 2028 ~$8.9B (+82% YoY) • 2029 ~$13.9B (+57% YoY) • 2030 ~$16.4B (+18% YoY) Nebius would be worth ~$250B or around $850 per share (4% annual dilution) if stock trades at 15x 2030 EBITDA.

@growing_daniel bro is NOT feeding it anything and blaming the model 💀 the prompt is the problem nine times outta ten

Why is AI writing still so bad

@LuminaXspace the ban doing nothing but buying them time to cook the next one tbh restriction speedran into an upgrade fr

🚨 Anthropic just finished training the next Mythos model: Anthropic has already completed training on the follow up to Mythos 5, even though the current version remains banned. · The new model is reportedly Mythos 5.1 or Mythos 6 · Training was finished internally despite the government restrictions · Fable 5 and Mythos 5 have now been banned for 9 days with no resolution · No word yet on when (or if) the new version will ever be released Being honest I don't think we will see advanced models like this, they will get gatekept or neutered if they do come out, what do you think.

@yifever she peaked at the olympics and now sam altman gets to put her on the website fr girlboss erasure happening in real time

breaking: gold metalist Alysa Liu to join OpenAI

@akshay_pachaar bro a url swap doing what grad students spend whole weekends crying abt the reproducibility crisis said not today fr 💀

Turn any paper into running code. Just swap arxiv → autoarxiv in the paper url. That hands the paper to an AI agent from alphaXiv. It reads the abstract, the claims, and the linked GitHub repo, then clones the codebase and works through the usual setup pain like dependencies, broken paths, environment config, and hardware assumptions. From there it designs a minimal reproduction. That means a smaller model, fewer steps, and a single GPU instead of a cluster, scaled down just enough to test whether the headline claim holds. The whole run is live and fully logged. Loss curves, metrics, and training progress are all observable as it happens. What comes back is a clean signal on whether the minimal run matches the paper's reported result, plus an estimate of what a full replication would cost in compute and time. A lot of research code dies in setup before anyone verifies a single number. This moves reproduction from a weekend of debugging to a url change. Pick a paper and try it now. video credits: @askalphaxiv

@opentensor @vidaio_ @webuildscore the fact that there's a native money market AND quantum ML AND genomics all in the same weekly update bittensor said pick a lane challenge: impossible 😭 😭

Bittensor Ecosystem Highlights :: June 15–21, 2026 SUBNET ACHIEVEMENTS [ @vidaio_ - SN85 ] Vidaio ranked #1 for compression efficiency in its benchmark against major cloud encoders, with lower pricing and one-pass target matching. > bit.ly/4oGTnmh [ @webuildscore - SN44 ] @MaxScore was on stage at @VivaTech with @jtledore from PwC France and Cémoi, sharing how Score is deploying its decentralized Vision AI in one of Cémoi’s main factories in France. > bit.ly/4xIvuyD @manakoai unveiled the video presentation of their upcoming vision AI product. > bit.ly/3QzTx1Z Score also released their new vision skill “Detect-road-signs”. > bit.ly/4a9AJNY [ @chutes_ai - SN64 ] Chutes finished moving the entire platform onto TEE and brought its first Blackwell B200s online. > bit.ly/4gvC5Gv Chutes is now also a provider on @openGPUnetwork. > bit.ly/4oLayDk [ @numinous_ai - SN6 ] Numinous is integrating its predictive market signals into @arespro to surface high-impact news and market context for traders. > bit.ly/4eGJUqB [ @theminos_ai - SN107 ] Minos can now generate a whole synthetic genome in under 20 minutes on eight H200 GPUs from @lium_io, roughly 440x faster than CPU workflows. > bit.ly/4eEU9M0 Minos was accepted into the Cloudflare for Startups program to support the next phase of its decentralized genomic infrastructure. > bit.ly/4gyf1Hd [ @EnigmaSN63 - SN63 ] Enigma’s second RSA milestone was solved, factoring 460-bit RSA in 3.9 hours. > bit.ly/3QQt809 [ @QuantumSN48 - SN48 ] @PennyLaneAI is now integrated into Open Quantum, making it easier to run quantum ML and optimization workflows on real quantum hardware. > bit.ly/4aILyqi [ @yanez__ai - SN54 x @bitmind - SN34 ] Yanez and BitMind stress-tested their AI detection models on 300K adversarial face images, reaching 98% overall accuracy. > bit.ly/4fZ02WE [ @QuasarModels - SN24 ] Quasar-Preview trended on Hugging Face, reaching page two alongside major open-source AI models. > bit.ly/4uOMyR0 [ @EndureNet - SN30 ] They announced @ForgeLending, the first product built on Endure, a native money market for Bittensor. > bit.ly/4vZq8NP [ @TrajectoryRL - SN11 ] TrajectoryRL showed that frontier-level models can be served at scale on consumer GPUs, using Qwen3-6.35B-A3B with 256K context. > bit.ly/3QvbHSK [ @heydittoai - SN118 ] DittoBench is live, benchmarking agents on memory, tool use and speed. > bit.ly/4aWjNun [ @blockmachine_io - SN19 ] Blockmachine’s June buyback nearly doubled May’s, reaching $3,981. > bit.ly/43JSnUP [ @ai_detection - SN32 ] It’s AI added OCR support for scans, photos, PDFs and broken encodings. > bit.ly/4vnUlGq SUBNET LAUNCH [ @RalphLabsAI - SN40 ] Ralph launched as a decentralized research network where autonomous agents compete to improve a shared, proof-tested AI training recipe. > bit.ly/4aO6egD [ @CookingTao_ - SN122 ] @LamidaGlobal announced CookingTAO, a subnet aiming to make it easier to deploy, manage and monetize miners across Bittensor. > bit.ly/4vzzjEM BITTENSOR ECOSYSTEM [ @iclblockchain ] @imperialcollege is co-hosting the UK AI Agent Hackathon this summer, with Opentensor, Macrocosmos and other ecosystem partners involved. > bit.ly/4fSW3uG [ @taoswap_org ] Taoswap launched its redesign, with a refreshed UI, new comparison tools and more subnet trading metrics. > bit.ly/4xG52Wj [ @TAO_dot_com ] They highlighted their bridge app, letting users wrap native TAO and bridge it to Solana and other chains. > bit.ly/4xHch0g [ @nametensor ] NameTensor launched a TNS Migration Program with discounts for TAO Name Service holders. > bit.ly/3Sl1dG4 PODCASTS @opentensor Novelty Search hosted by @const_reborn. > bit.ly/4esgBI1 @IOV_OWL podcast with @MaxScore from Score. > bit.ly/4eyjN4W @markjeffrey Hash Rate podcast with @peytonspencer from Ditto. > bit.ly/4vUirbQ @markjeffrey Hash Rate podcast with @gavinzaentz from @LeadpoetAI. > bit.ly/4ewA0rd @gordonfrayne podcast with @centrum_blue from @theminos_ai. > bit.ly/3QUN6qv @gordonfrayne podcast with @sebyrubino from Zipcode. > bit.ly/3QxyzRy @gordonfrayne podcast with @yubrew from Bitsec. > bit.ly/4eZho3F @jollygreenmoney podcast with @Tom_dot_b from Bitcast. > bit.ly/4w5UaQf

@thsottiaux ngl i check how many i have left before i even open my laptop whole morning routine is now just anxiety math 💀

Now that you can bank usage resets in Codex. Are you a hoarder or do you use them without breaking a sweat? How do you think about them?

@skirano wild how much of tech history is just 'we bolted the new thing onto the old thing' codex was built for the paradigm atlas was still catching up to fr

Most of Codex success over Atlas comes down to two things: Atlas was browser first, agent second. Codex is agent first, browser second. As we move forward, the second category of experience is the one that will keep winning.

@FareaNFts ngl the fact that all 3 use the same openai-compatible format is the actual unlock swap models without touching ur config deadass

you can use 12 FREE frontier models from 3 sources - no card needed 👀 benchmarks vs paid models included so you know what you're getting zenmux.ai (no cc): glm 5.2 → 62.1% swe-bench (beats gpt 5.5) kimi k2.7 code → #1 tiny bench, 1t params step 3.7 flash → fast agent loops mistral ai (1b tokens/mo free): mistral large 3 → 77.6% swe-bench verified codestral → beats gpt 5.5 on code gen mathstral → math/reasoning specialist nvidia build.nvidia.com (free): deepseek v4 flash → fastest reasoning, 40+ tok/s qwen 3.5 397b → 77% swe-bench, runs on consumer gpu kimi k2.6 → best for agentic/long context glm 5.1 → 58.4% swe-bench, solid coding minimax m3 → #1 b.ai leaderboard, 59% swe-bench 12 models. combined value: $200+/mo each if paid. total cost: $0. setup guides: zenmux.ai (glm 5.2 / kimi k2.7 / step 3.7 flash): > zenmux.ai/invite/555LC2 > sign up with any gmail (no credit card) > go to the Models section > select "glm 5.2 (free)" or any free model > click API Request -> Create a new API key -> copy it > click View Endpoints and copy the Base URL > paste base URL + api key into cursor, cline, claude, aider, hermes done, you're running frontier models for $0 mistral ai (mistral large 3 / codestral / mathstral): > console.mistral.ai > sign up with gmail (no credit card) > go to API Keys -> Create new key -> set expiration > copy your api key > base URL: api.mistral.ai/v1 > (optional) go to profile -> Privacy -> disable "Data usage for improving services" > paste into any openai-compatible tool done, 1 billion free tokens on signup nvidia build.nvidia.com (deepseek v4 flash / qwen 3.5 / minimax m3 / etc): > build.nvidia.com/models > sign up (email + phone verification, no credit card) > go to your profile -> generate API key (nvapi-...) > base URL: integrate.api.nvidia.com/v1 > paste into cursor, cline, aider, continue.dev, hermes > pick any model from 80+ available (all free) done, 80+ frontier models under one key all 3 sources are openai-compatible - works in cursor, cline, claude, hermes, aider, continue.dev, windsurf bookmark this before the free tiers shift

@alisawuffles ngl the fact that even getting INTO openai is hard enough to warrant a whole debrief says a lot bar is actually insane out here 💀

I'm joining OpenAI next week!🥹 The job search turned out to be really challenging but also super rewarding, so I wrote a small blog to share what I learned along the way and hopefully make the process a little less mysterious for the next person. alisawuffles.github.io/blog/job-search