- Published on
科技推特精选 - 2026年6月4日
- Authors
- Name
- geeknotes
2026年6月4日 科技每日简报
Today's top tech conversations are led by @vllm_project, whose post about 'Congrats to the @googlegemma t...' garnered the highest engagement. Key themes trending across the top stories include https, model, today, capital, billion. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.
1. vllm_project (Group Score: 342.0 | Individual: 34.2)
Cluster: 15 tweets | Engagement: 211 (Avg: 166) | Type: Tech
Congrats to the @googlegemma team on the Gemma 4 12B launch 🎉 Day-0 support on vLLM is ready to go.
It's an encoder-free unified multimodal model — text, image, audio, and video all project straight into the LLM's embedding space, no separate vision or audio towers. 256K context, built-in thinking, native tool calling.
Reasoning + tool parsers (gemma4), vision, and audio all served through the OpenAI-compatible API.
🔗 Recipe: https://t.co/MGJcoQkwzz\n\nQT @googlegemma: Meet Gemma 4 12B!
A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license.
Bridging the gap between edge efficiency and advanced reasoning. Here is what’s new with Gemma 4 12B: 👇 https://t.co/gf4FZv0WZb
See 14 related tweets
- @lmsysorg: 🎉 Meet Gemma 4 12B Unified from @googlegemma! This is a 12B dense, encoder-free multimodal that runs...
- @clmt: RT @Google: Today we’re introducing Gemma 4 12B — our latest open model that brings advanced agentic...
- @kimmonismus: Gemma 4 12B shipped today under the label "encoder-free."
A local 12b model that shows really good...
- @victormustar: wow 👀\n\nQT @mervenoyann: Google dropped Gemma-4 12B, it's a beast 🔥
> unified: audio + image go...
- @tayarndt: I think you can also run it as low as eight gigs of ram. I may need to pull up my Neo.\n\nQT @vllm_p...
2. chhddavid (Group Score: 261.2 | Individual: 57.5)
Cluster: 8 tweets | Engagement: 828 (Avg: 29) | Type: Tech
RT @ideogram_ai: Introducing Ideogram 4.0: the best open image model in the world.
Think it. Make it. Own it.
Download the weights, fine-tune on your own data, and run it on your hardware. Live on every Ideogram plan and the API today. https://t.co/AdH9hfSEdb
See 7 related tweets
- @huggingface: Ideogram just released their latest and best v4 image model open weights
State of the art and open ...
- @a16z: Ideogram 4.0 is incredible.
And it's open. Download the weights below. https://t.co/gh07JxAB6m\n\nQ...
- @Redpoint: Congrats to @mo_norouzi and the @ideogram_ai team on their new model: Ideogram 4.0! This is now the ...
- @martin_casado: New leading open source image model!! Performance on design Arena is incredible.
So great to see t...
- @ComfyUI: Fresh Open weight model drop😋\n\nQT @ideogram_ai: Introducing Ideogram 4.0: the best open image mode...
3. StockMKTNewz (Group Score: 223.7 | Individual: 34.3)
Cluster: 10 tweets | Engagement: 256 (Avg: 421) | Type: Tech
Google 84.75 Billion from GOOGL JUST ANNOUNCED AN $80 BILLION CAPITAL RAISE TO BUILD AI INFRASTRUCTURE
And Berkshire Hathaway 10 billion check to get in.
Here's the full breakdown:
THE DEAL:
- $30B in underwritten public offerings
- $40B through an at-the-market stock program starting Q3 2026
- $10B private placement to Berkshire Hathaway
THE BERKSHIRE PIECE:
- 351.81 per share
- 348.20 per share
- Berkshire has been building this position since Q3 2025
THE PURPOSE:
- Scale AI compute infrastructure to meet "unprecedented customer demand"
- Approximately $30B of the ATM proceeds will cover 2026 employee equity tax obligations
- Remaining proceeds go directly to AI infrastructure buildout
See 9 related tweets
- @StockSavvyShay: 30B of its $80B equity raise to cover tax obligations tied to employee stock a...
- @StockSavvyShay: 85B from $80B after its convertible stock sale was reportedly ...
- @edzitron: This is the goofiest way to look at this imaginable, I’m sorry man. It’s a sign that Google and by e...
- @business: Google’s doing a massive multi-billion dollar secondary equity offering and has Berkshire Hathaway s...
- @Techmeme: Google selling $10B of equity to Berkshire is fascinating and functions as a signal that AI demand i...
4. VaibhavSisinty (Group Score: 188.6 | Individual: 33.5)
Cluster: 7 tweets | Engagement: 66 (Avg: 63) | Type: Tech
Man. Hermes Agent just dropped something insane. 🤯
140K GitHub stars in 3 months. Most used agent on OpenRouter. And now they've shipped a full desktop app for Mac, Windows, and Linux.
No terminal. No setup. Just download and go.
The thing that sets it apart is dead simple. One agent. One memory.
Everywhere.
Slack. WhatsApp. Telegram. Discord. Signal. Email. Desktop. It doesn't matter where you talk to it. It remembers everything across every surface.
Schedules tasks in plain English. Runs subagents for complex work.
Everything stays local on your machine. Nothing leaves.
Open-source. MIT license. Free.\n\nQT @NousResearch: The next evolution of Hermes Agent is here!
Introducing Hermes Desktop: everything you love about Hermes, now native on your machine.
First demoed in Jensen's GTC keynote, it's now in public preview. https://t.co/8ND1k8hyaz
See 6 related tweets
- @Marktechpost: This is really Big update from Nous Research: Hermes Desktop is now in public preview.
If you’ve fo...
- @TeksEdge: 😮 Biggest 🤖 Agent News of the Day! 🏛️Hermes Desktop makes running @NousResearch Hermes Agent(s) easy...
- @WesRoth: Nous Research launched Hermes Desktop, bringing Hermes Agent into a native desktop app for local mac...
- @akshay_pachaar: Hermes agent just left the terminal.
𝗛𝗲𝗿𝗺𝗲𝘀 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 dropped yesterday. native app for macOS, Window...
- @LightningAI: Hermes Agent crossed 140,000 GitHub stars in under three months and is now the most-used agent on Op...
5. omarsar0 (Group Score: 171.2 | Individual: 29.7)
Cluster: 8 tweets | Engagement: 77 (Avg: 104) | Type: Tech
Another banger open-source release.
Miso One is an 8B text-to-speech model with real emotional range, so voiceovers carry warmth, hesitation, and excitement instead of sounding flat.
It's purpose-built for voiceover work like shorts, podcasts, and educational content, and it runs at 110ms latency, which is faster than human reaction time.
The best part is that the weights are fully open source, so you can clone the repo, self-host, fine-tune, and keep your data private.
Worth checking out if you're building voice into your tools and products: https://t.co/1gFeKGyxGm\n\nQT @AodenTeoMT: Today, we’re excited to introduce Miso One, the most emotive voice model in the world.
Miso One is an 8-billion-parameter text-to-speech model for highly expressive speech generation. It emotes like a human and responds faster than a human, with just 110 milliseconds of latency.
We’ve open-sourced the model weights, with API access coming soon.
Hear how Miso One sounds in the thread below.
See 7 related tweets
- @kimmonismus: Miso One is live: an open-weights voice model built to sound like a real person reading, with actual...
- @ai_explorer25: My elevenlabs bill last month: embarrassingly high
The revenue from the videos that actually used i...
- @dair_ai: RT @omarsar0: Another banger open-source release.
Miso One is an 8B text-to-speech model with real...
- @FellMentKE: Most TTS models benchmark well and sound completely dead in production.
Miso One is different, it c...
- @ycombinator: RT @AodenTeoMT: Today, we’re excited to introduce Miso One, the most emotive voice model in the worl...
6. shiri_shh (Group Score: 169.3 | Individual: 32.7)
Cluster: 8 tweets | Engagement: 37 (Avg: 198) | Type: Tech
Why is no one paying attention to AI Music?
Suno is generating 7 MILLION songs per day.
That’s basically Spotify’s entire catalog every two weeks 😭
already at $300M+ ARR with 2 MILLION paid users
and raised another 5.4B valuation today https://t.co/ePbkhFIB5n\n\nQT @lightspeedvp: We’re tripling down on our investment in @suno.
After leading Suno’s Series B in 2024, we're proud to invest again in their 5.4B.
Suno is a company that allows anyone, regardless of skill level, to create music. The AI platform was built to increase access to music creation. Now, videos using songs made with Suno are going viral.
AI has the potential to make a major impact on accessibility to the arts. Suno is a great example of this, offering users a platform to express themselves through music.
Congratulations to the Suno team on their growth! We believe the platform has big potential, and are looking forward to seeing how Suno will continue to progress.
CC: @bsomaia
See 7 related tweets
@business: Suno, a startup that enables people to make music using AI, has raised $400 million at a valuation o...
@mmurph: .@suno is an absolute phenomenon that continues to surprise on the upside! love companies that are l...
@moneycontrolcom: AI music startup Suno raises over 5.4 billion valuation
Suno raises $400M, doubl...
@MenloVentures: Thrilled to double down on @suno's $400M round, after leading its Series C last fall!
Suno is makin...
- @TechCrunch: Still facing copyright lawsuits, AI music generator Suno raises another $400M https://t.co/bgupVM7Hl...
7. gdb (Group Score: 163.8 | Individual: 30.8)
Cluster: 6 tweets | Engagement: 918 (Avg: 817) | Type: Tech
codex for computer work is growing very fast\n\nQT @OpenAINewsroom: Codex now has more than 5M weekly active users.
But the bigger story is what people are using it for: not just writing code, but getting more work done across research, analysis, content, and operations.
Our new report on how Codex is becoming a productivity tool for knowledge work: https://t.co/XZMPR9cEge
See 5 related tweets
- @WesRoth: Codex now has more than 5 million weekly active users, up more than 6x since the desktop app launche...
- @RocM301: Codex直接上线6款职场专属插件，数据分析师、营销、设计师、投行人都能直接用，62个常用App+110项技能打包好。\n\nQT @OpenAI: We’re making Codex more u...
- @rohanpaul_ai: RT @rohanpaul_ai: OpenAI just gave Codex a major upgrade.
From a coding assistant into a workspace ...
- @steph_palazzolo: Codex is making OpenAI engineers a lot more productive—but, that productivity is a double-edged swor...
- @ajambrosino: not to worry– we're also taking Codex to the next level for software development. what should we foc...
8. pmarca (Group Score: 154.8 | Individual: 32.0)
Cluster: 7 tweets | Engagement: 237 (Avg: 736) | Type: Tech
Interesting.\n\nQT @TMTLongShort: The equation is fairly straightforward:
Competent employees x AI tokens = Accelerating business & market share gain
Incompetent employees x AI tokens = slop
Companies are now realizing they have a lot of shitty employees.
They aren’t going to permanently cut spend on tokens. They can’t afford to because of game theory.
So instead they will fire the employees they believe are incompetent to make room for higher token budgets for those that are competent.
Lots of orgs however have a managerial class that doesn’t optimize for share gain and winning in general.
Thats fine. A wave of startups and existing platforms who can effectively leverage AI to expand scope of their business will crush the incompetent at a rate that will leave analysts and managers dizzy.
Change is coming. Fast. And reflexively the faster the change the higher the panic the lower the ROI threshold the more revenue and capital accrues to the labs the faster the models improve. And so on.
See 6 related tweets
- @pmarca: RT @TMTLongShort: The equation is fairly straightforward:
Competent employees x AI tokens = Accele...
- @JayminSOfficial: Many people are focused on @DavidSolomon's "greed" comment.
What matters more is that markets are t...
- @koylanai: Build a simple agent that scans your engineers’ Cursor, Codex, and Claude Code logs every day.
It q...
- @Yuchenj_UW: More and more engineers are now burning more money on AI tokens than their base salaries.
Tech comp...
- @chris_j_paxton: "People love using AI because it makes their jobs easier and more fun" is not inconsistent with "the...
9. FellMentKE (Group Score: 151.7 | Individual: 35.0)
Cluster: 5 tweets | Engagement: 108 (Avg: 128) | Type: Tech
Building email flows has always been the worst part of growth.
Nitrosend letting you one shot entire campaigns from Claude with pixel perfect design is the hostage rescue we needed.
Try Nitrosend free: https://t.co/g1fhOqiohm\n\nQT @gthartley: I just sent our launch announcement to 10,000 people.
It took one prompt in Claude.
Today we're launching @nitrosendx - the first email platform with no dashboard.
→ https://t.co/AwAv8mP5KM https://t.co/quW4W0JPCG
See 4 related tweets
- @Scobleizer: Founders that do email campaigns, this one is for you.
Have a project to market my AI news site: ht...
- @ai_explorer25: Honestly the AI isn't even my favorite part of Nitrosend. Everyone's got AI these days
It's that un...
- @cgtwts: email marketing just became a prompt.
“Email for Claude” just launched.
It writes the copy, design...
- @shiri_shh: "Email for Claude" just launched. The first AI-native email platform built to be controlled from C...
10. dee_bosa (Group Score: 150.3 | Individual: 25.5)
Cluster: 9 tweets | Engagement: 88 (Avg: 143) | Type: Tech
Uber might look cheap here but budget discipline around AI is probably about to become normal
Unlimited token burn isn't a strategy and being disciplined about usage doesn't seem insane\n\nQT @HighyieldHarry: Capping people at $1,500/month in AI tool spend tells me you don't who the AI power users are and who is checking the weather.
See 8 related tweets
- @MilkRoadAI: RT @WhiteCollarExit: The next AI trade may start with the UBER.
Reportedly, ...
- @alex: RT @Techmeme: Uber says it has limited all employees to $1,500 in monthly token spending per AI codi...
- @garrytan: Time to short Uber and long DoorDash\n\nQT @negligible_cap: *UBER SETS $1,500 MONTHLY CAP ON SOME AI...
- @dee_bosa: the 3 phases of AI sticker shock:
- board asks “whats our AI strategy?”
- company says “everyone u...
- @DataChaz: BREAKING: @Uber is reportedly capping employee use of AI vibe-coding tools at $1,500 per month after...
11. business (Group Score: 142.1 | Individual: 29.8)
Cluster: 6 tweets | Engagement: 53 (Avg: 72) | Type: Tech
Meta Platforms Inc. is selling businesses access to an artificial intelligence agent for the first time, its latest effort to generate revenue to offset the company’s hefty AI investments. https://t.co/YTJDKiBISv
See 5 related tweets
- @CNBC: Meta’s effort to push beyond ads and to make money from artificial intelligence is now targeting the...
- @WSJ: For 20-plus years, Meta Platforms has been a consumer business. That is now starting to change.
Th...
- @StockMKTNewz: ZUCKERBERG IS NOW TRYING TO SELL AI TO BUSINESSES
Meta Platforms $META is selling businesses access...
- @Reuters: Meta enters enterprise AI race with new business agent https://t.co/INGXbfX4TT https://t.co/INGXbfX4...
- @WSJTech: Mark Zuckerberg wants Meta Platforms’ new AI agents to run your whole business https://t.co/i0f8CW1S...
12. peterwildeford (Group Score: 135.1 | Individual: 27.0)
Cluster: 7 tweets | Engagement: 21 (Avg: 130) | Type: Tech
OPENAI: "policymakers should require the most capable frontier models to undergo a CAISI evaluation before public release"\n\nQT @OpenAINewsroom: There’s real momentum right now for AI safety policy. Yesterday’s EO on cyber was an important step forward.
We’re proposing a set of ideas for policymakers to consider next and to put the US out in front on frontier safety.
See 6 related tweets
- @peterwildeford: RT @_NathanCalvin: Overall my reaction to this doc is pleasant surprise, and I think this proposal t...
- @peterwildeford: OPENAI: "We also see early signs of recursive self-improvement in today's systems". RSI is "potentia...
- @mark_k: Me when I hear "AI Safety":
🤮\n\nQT @OpenAINewsroom: There’s real momentum right now for AI safety ...
- @deanwball: This seems reasonable!
Having CAISI—a civilian agency—conduct this testing in primarily non-classif...
- @OpenAINewsroom: There’s real momentum right now for AI safety policy. Yesterday’s EO on cyber was an important step ...
13. wallstengine (Group Score: 133.3 | Individual: 42.6)
Cluster: 4 tweets | Engagement: 351 (Avg: 120) | Type: Tech
$AVGO Q2’26 EARNINGS HIGHLIGHTS
🔹 Revenue: 22.13B) 🟢; +48% YoY 🔹 Adj. EPS: 2.39) 🟢; +54% YoY 🔹 Semiconductor Solutions: 14.65B) 🟢; +79% YoY 🔹 AI Semiconductor: 11.3B) 🟡; +143% YoY
Q3 Guide: 🔹 Revenue: ~28B; buyside 16B (buyside Est. ~$18B) 🟡 🔹 Non-GAAP Operating Income: ~67% of projected revenue 🔹 Adjusted EBITDA: ~68% of projected revenue
Segment Performance: 🔹 Semiconductor Solutions Revenue: 7.178B; +9% YoY
Financials: 🔹 FCF: 12.074B; +55% YoY 🔹 Adjusted EBITDA: 10.493B; +60% YoY 🔹 Cash & Cash Equivalents: $19.628B
Capital Return: 🔹 Dividend: 3.092B
Commentary: 🔸 “Broadcom achieved record revenue, operating profit and free cash flow in Q2 driven by accelerating growth in AI semiconductor revenue and strong operating leverage.”
🔸 “Q2 semiconductor revenue from AI of $10.8 billion grew 143% year-over-year, above our forecast, driven by increasing demand for custom AI accelerators and AI networking.”
🔸 “The momentum continues and in Q3 we expect semiconductor revenue from AI to grow over 200 percent year-over-year to $16.0 billion.”
🔸 “In Q3 we expect consolidated revenue growth to increase 84% year-over-year to $29.4 billion, with non-GAAP operating margin stable at 67% reflecting our strong operating leverage.”
🔹 Data conflict: Revenue shows 22.187B in the press release/Q2 earnings details
See 3 related tweets
- @wallstengine: $CRWD Q1’27 EARNINGS HIGHLIGHTS
🔹 Revenue: 1.38B) 🟡; +26% YoY 🔹 Adj. EPS: $1.10 (Est....
- @wallstengine: $TLYS Q1’26 EARNINGS HIGHLIGHTS
🔹 EPS: -0.33 Est. 🟢 🔹 Revenue: $124.718M; +15.9% YoY 🔹 C...
- @StockSavvyShay: $AVGO Q2 EARNINGS
• Revenue 22.1B • EPS 2.39 • EBITDA $15.2B vs Est....
14. ajambrosino (Group Score: 131.0 | Individual: 35.6)
Cluster: 4 tweets | Engagement: 986 (Avg: 360) | Type: Tech
a few months ago in an interview, someone asked me where we wanted to take the codex app. i answered with something along the lines of: we intend to make it the best app ever built for desktop, full stop.
it probably felt like a little much, but i meant it. and it feels less potentially hyperbolic now than it did then.
Codex is the best way to build software. it is now the best way to do many other things too. we will lean into both, and much of it will mean we blend a lot with ChatGPT. we will do this only when we can deliver something incredible and better than what the two separate current things can deliver alone.
we will also combine the best parts of cloud and local environments (and yes, windows+linux). and the best of instant responses and long-running objectives, like /goal. we will do so with Taste™
outside of the incredible gpt models, 3 things have made the Codex app what it is:
- an opinionated view of how agents should work with a high quality bar
- a tight and honest dogfooding loop
- you
those 3 things will continue to be P0. LFG.
See 3 related tweets
- @dkundel: Today it's been 5 months since we launched the codex app. it's wild how much things have changed in ...
- @SIGKITTEN: im still new (2 days in) and i can still put my cynical kitty glasses on.
A small part of me was ho...
- @reach_vb: When I joined OpenAI, the codex app was still in its early days of being put together and being a st...
15. a16z (Group Score: 124.1 | Individual: 43.5)
Cluster: 4 tweets | Engagement: 791 (Avg: 251) | Type: Tech
We’re thrilled to lead Town’s $55M Series A.
Town is a personal AI assistant that works across the tools you already use - email, calendar, Slack, docs, WhatsApp, desktop, web. It learns how you work and starts proactively pitching in.
People are already leaning on Town for the kind of work that’s personal and operationally messy: running recruiting pipelines, juggling school logistics, processing handwritten grant requests, prepping summaries, drafting follow-ups, catching the stuff that would otherwise slip.
The longer you use it, the more it picks up: your voice, your relationships, your preferences, your routines, what you actually care about.
Jean-Denis Greze and Tony Vincent are the right team for something this hard. JDG was CTO at Plaid and an engineering leader at Dropbox. Tony led product and AI at Google and design at Dropbox.
Welcome, JDG, Tony, and the Town team, to the a16z family.I
By @arampell and @venturetwins\n\nQT @jgreze: Today, we’re launching @TownAI: the AI assistant that learns you.
We’re coming out of beta with a $55M Series A led by @ARampell at @a16z, with participation from @KirstenGreen at @forerunnervc and continued support from @firstround, @altcap, and @conviction.
Right now, getting real value from AI means prompting, configuring, building workflows, managing agents.
We think that’s backwards.
The future of AI is a companion that already knows you and how you work. Town connects across your inbox, calendar, Slack, docs, messages, and workflows to understand what you need, then starts doing the work with you.
Drafting. Scheduling. Project tracking. Follow-ups. Context gathering. Multi-step tasks. And it only acts when you say so.
All adapting to your voice, priorities, routines, and relationships over time.
Your Townie is the AI assistant you actually need.
See 3 related tweets
- @saranormous: Congrats to team @TownAI! One of the first viscerally different AI experiences. Long horizon agents,...
- @arampell: RT @jgreze: Today, we’re launching @TownAI: the AI assistant that learns you.
We’re coming out of b...
- @jamdac: RT @a16z: We’re thrilled to lead Town’s $55M Series A.
Town is a personal AI assistant that works a...
16. ShanuMathew93 (Group Score: 120.7 | Individual: 33.4)
Cluster: 4 tweets | Engagement: 20 (Avg: 30) | Type: Tech
So I agree it’s been the company that has acted on AI the most thus far in its recommendation systems and actually defended that point in a recent post
But as a product user, I feel like my generation who grew up hooked on the META ecosystem is using the products less and less
Might be a function of demographics (early 30s) but it is notable\n\nQT @WOLF_Financial: JENSEN HUANG WAS ASKED WHO IS USING AI BETTER THAN ANYONE ELSE IN THE WORLD
His answer wasn't OpenAI. It wasn't Google. It wasn't even Nvidia itself.
Here's what he said on CNBC:
"Nobody uses AI better than Meta."
He explained that Meta went from a classical recommender system running on CPUs to a generative AI agentic system making recommendations across the entire platform.
"Everything from the way social media works and the way they recommend ads and help advertisers create content has fundamentally been changed. And their earnings show it."
"That's the reason why they're investing so hard. They see a much larger future potential for it."
Mark Zuckerberg 70B+ on AI this year.
See 3 related tweets
- @StockMKTNewz: Inject it in my veins\n\nQT @WOLF_Financial: JENSEN HUANG WAS ASKED WHO IS USING AI BETTER THAN ANYO...
- @MilkRoadAI: RT @MilkRoadAI: Meta is extremely UNDERVALUED and Jensen Huang just explained exactly why the market...
- @ShanuMathew93: They’ve probably been the biggest beneficiary of AI to improve recommendation engines to date…?
And...
17. BullTheoryio (Group Score: 120.7 | Individual: 33.6)
Cluster: 5 tweets | Engagement: 398 (Avg: 1210) | Type: Tech
BREAKING: Ray Dalio just said the AI market is a bubble and it will burst.
"All great technology changes produce bubbles," Dalio told Bloomberg. "The pricking is the converting of wealth into money" right now, every major tech company is pouring hundreds of billions into AI infrastructure and booking it as investment.
The moment investors demand actual returns, companies will have to show that the money spent is generating real profits from real customers. If the revenue is not there, valuations collapse and right now, the revenue is not there.
AI companies are spending 60 billion annually on cloud infrastructure against $25 billion in actual revenue.
Less than 1% of executives globally report meaningful ROI from their AI investments. 95% of enterprise AI pilots have failed to deliver measurable returns according to MIT.
The entire $2 trillion cloud backlog held by Microsoft, Oracle, Google, and Amazon is anchored by two unprofitable companies: OpenAI and Anthropic.
By 2030, the industry needs 800 billion short.
Dalio is not saying the technology is fake. He is saying the economics do not work yet and every bubble in history has ended the same way when that moment of reckoning arrived.
See 4 related tweets
- @cryptorover: 🚨 THERE IS VERY LITTLE DIFFERENCE BETWEEN THE AI BUBBLE AND THE DOT-COM BUBBLE.
Microsoft invested ...
- @StockMKTNewz: - RAY DALIO SAYS AI BUBBLE WILL BURST EVENTUALLY https://t.co/5PqIE1H4C0\n\nQT @WOLF_Financial: RAY ...
- @Cointelegraph: 🚨 INSIGHT: Billionaire investor Ray Dalio warned that the booming AI market is showing signs of a bu...
- @edzitron: RT @edzitron: Went on Bloomberg - Anthropic and OpenAI are dangerous and unsustainable companies tha...
18. rohanpaul_ai (Group Score: 115.6 | Individual: 31.7)
Cluster: 4 tweets | Engagement: 2 (Avg: 62) | Type: Tech
In the world of finance the biggest challenge with AI isn't intelligence. It's trust.
Investment teams are still losing too much time turning messy deal files & excels into reports they can actually trust.
These guys just launched Leni, a multi-agent system for serious work in real estate, private equity, and investment finance.
Claims 80% faster reports, beats GPT/Claude on investment tasks, and full source-traceable outputs.
It is designed to run longer tasks, cross-check itself, and return finished deliverables like underwriting workbooks, market studies, IC memos, and reporting\n\nQT @arunabh_D: "Mostly right" is useless when millions of dollars are on the line.
That's why we built Leni: the world's most accurate AI for investment work. It analyzes hundreds of files at once, reasons through complex problems, validates its outputs, and delivers finished work instead of making users babysit prompts.
Over the last few years, Leni has been used by institutions to manage over $80B in assets, and we have learned that the biggest challenge with AI isn't intelligence.
It's trust.
Many AI systems sound convincing. Few can be trusted for serious investment decisions.
So we aimed to build one that is accurate, verifiable, and dependable.
Everything we've learned from processing over 100 million rows of data, building our own verification models, evaluation frameworks, and investment workflows, is now embedded into a powerful self-serve platform designed to help investors make institutional-grade decisions with confidence.
Built for serious work.
Built for accuracy.
Try Leni today and experience the difference. 🚀
See 3 related tweets
- @godofprompt: AI demos are cheap. Trust is expensive.
In finance, one confident wrong answer can burn more than y...
- @heynavtoor: your team uses Claude. your competitor uses Claude. where's the edge?
Leni just launched an AI buil...
- @alex_prompter: Hallucinations are funny until they hit a balance sheet.
Everyone wants AI speed. Nobody wants to b...
19. JamesMontemagno (Group Score: 112.2 | Individual: 42.5)
Cluster: 3 tweets | Engagement: 992 (Avg: 204) | Type: Tech
RT @mustafasuleyman: Super excited to announce seven new world-class MAI models today. They represent what we consider a new era in AI designed to keep you in control and on the frontier. First is our text foundation model, MAI-Thinking-1, exceptionally strong on reasoning and SWE tasks.
- It’s a 35B active parameter MoE with a 256K context window. Independent human raters on Surge prefer it for overall quality in blind side-by-sides versus Sonnet 4.6, and it’s achieved 97% on AIME 2025, the key measure of its general-purpose reasoning abilities.
- It's at 53% on SWE Bench Pro, placing it right alongside Opus 4.6 on one of the toughest coding benchmarks.
- And since we co-designed our models with our own silicon, MAI-Thinking-1 is optimized on our MAIA 200 chip. Benchmarking head-to-head against the GB200, we see 30% better performance per dollar as well as a 1.4x performance-per-watt gain when running our MAI models on the MAIA 200 end-to-end.
Next is MAI-Image-2.5 and its Flash variant. Two super strong models now at #2 on the leaderboards, surpassing the score of Nano Banana 2 on image editing.
Last for now is MAI-Code-1-Flash, our new inference efficient coding model, especially tuned for VS Code and GitHub Copilot CLI.
- Code-1-Flash achieves 51% on SWE Bench Pro, despite having just 5B parameters, putting it closer to Haiku in size but cheaper in cost.
All of this is the foundation for Microsoft Frontier Tuning. It lets you customize our models to create custom, company-specific agents that only you control. You can make our model, your model. Your data. Your agents. Your moat.
Early adopters are already seeing a difference. When we tuned our models for McKinsey’s tasks, MAI delivered the highest win rate, outperforming GPT-5.5 on quality, while being 10x lower on cost.
Also really excited to be collaborating with the amazing team at Mayo Clinic to jointly train a new frontier AI model for healthcare.
Our announcements today mark another milestone on the road to humanist superintelligence. You can learn more and about our other new models in our latest blog: https://t.co/v65eop5Ixq
See 2 related tweets
- @yacineMTB: everyone who doubted Mustafa real quiet right now\n\nQT @mustafasuleyman: Super excited to announce ...
- @TheStalwart: I’m really curious about this. How much extra juice is there, performance wise, for model-specific s...
20. business (Group Score: 107.4 | Individual: 65.5)
Cluster: 3 tweets | Engagement: 752 (Avg: 72) | Type: Tech
India’s ambition to build and export a sovereign AI template across the world is colliding with structural constraints: years of underinvestment in compute capacity, a late start in building the most advanced AI models, deep reliance on foreign cloud providers and a venture capital culture wary of the vast, risky bets that define the global AI race. https://t.co/0HhK3Iwzqw
🎨: 731; 📷️: Getty Images (2)
See 2 related tweets
- @SarithaRai: Deep dive: India’s ambition to build and export a sovereign AI template across the world is collidin...
- @Techmeme: How India's ambitions to build and export a sovereign AI template are colliding with structural cons...