科技推文精选 - 2026年6月12日

科技每日简报 | 2026-06-12

Today's top tech conversations are led by @WSJ, whose post about 'OpenAI is considering drastic ...' garnered the highest engagement. Key themes trending across the top stories include anthropic, fable, agent, model, openai. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. WSJ (Group Score: 558.6 | Individual: 44.4)

Cluster: 31 tweets | Engagement: 1138 (Avg: 133) | Type: Tech

OpenAI is considering drastic price cuts as it seeks to win over customers from archrival Anthropic https://t.co/vwWcJJd1x8

See 30 related tweets

@mark_k: OpenAI is reportedly considering drastic token price cuts to pull customers away from Anthropic, per...
@negligible_cap: *OPENAI MULLS SIGNIFICANT CUTS TO WHAT IT CHARGES FOR TOKENS:WSJ

OpenAI cutting token prices to try...

@BullTheoryio: BREAKING: OpenAI Is Reportedly Considering Major AI Price Cuts as Competition Intensifies, according...
@michalmalewicz: We just need 18B people at lower prices to break even now.

Sam tomorrow: Hey guys, make babies plz...

@zerohedge: BOOM

*OpenAI Considers Drastic Price Cuts, Anticipating Costly War For Users With Anthropic -- WSJ...

2. WIRED (Group Score: 401.6 | Individual: 47.8)

Cluster: 17 tweets | Engagement: 396 (Avg: 34) | Type: Tech

RT @ZeffMax: NEW: Anthropic is walking back Claude Fable 5's policy to covertly degrade performance for competing AI researchers, after facing fierce backlash.

“We’re changing Fable 5’s safeguards for frontier LLM development to make them visible,” Anthropic tells WIRED. “We made the wrong tradeoff and we apologize for not getting the balance right.”

See 16 related tweets

@MTSlive: SITUATION UPDATE: Anthropic is reversing its Fable 5 policy of covertly degrading performance for co...
@chandrarsrikant: 🚨🚨After backlash, Anthropic to revise Claude Fable 5's AI development restrictions

IPO-bound Anthro...

@Scobleizer: Well, this was to be somewhat expected that a PR team would walk back a bad decision.

It doesn't se...

@AndrewCurran_: New policy: 'Starting this week, flagged requests will visibly fall back to Opus 4.8. On the API, an...
@altryne: Damn, look at that!

I was about to go and cover the whole "we degrade model performance silently" ...

3. MTSlive (Group Score: 336.8 | Individual: 52.0)

Cluster: 13 tweets | Engagement: 856 (Avg: 131) | Type: Tech

SITUATION BREWING: Prometheus, the industrial AI startup led by Jeff Bezos, is set to announce a $12B Series B at a$ 41B valuation, per Axios.

The company is building what it calls an “artificial general engineer.”

See 12 related tweets

@StockMKTNewz: Jeff Bezos' AI company is out of stealth at a $41 Billion valuation\n\nQT @IPONewsroom_: JEFF BEZOS ...
@IPONewsroom_: JEFF BEZOS JUST EMERGED FROM STEALTH WITH A $41 BILLION AI STARTUP CALLED PROMETHEUS

$12 billion ra...

@DeItaone: BEZOS SAYS AI WILL BOOST JOBS, LAUNCHES PROMETHEUS

Jeff Bezos said AI will create a labor shortage ...

@Techmeme: Jeff Bezos' Prometheus, which is building AI models for physical tasks, raised a $12B Series B at a ...
@Polymarket: NEW: Jeff Bezos reveals his new startup Prometheus is building an “artificial general engineer” as i...

4. IPONewsroom_ (Group Score: 263.0 | Individual: 31.7)

Cluster: 13 tweets | Engagement: 34 (Avg: 146) | Type: Tech

JUST IN: OPENAI HAS AGREED TO ACQUIRE ONA, A CLOUD STARTUP THAT POWERS AI AGENTS

The team is being folded into OpenAI's Codex coding effort, and one stat from the announcement explains why OpenAI is moving fast:

Codex now has more than 5 million weekly users.

Ona's platform provides "secure, persistent environments where agents can access the tools, systems, and context they need to make progress over time," per OpenAI.

The deal has not closed yet.

OpenAI and Anthropic have both filed confidential paperwork to go public, with debuts targeted as soon as this fall, per Bloomberg.

See 12 related tweets

@business: OpenAI has agreed to acquire Ona, a startup that offers cloud services to support artificial intelli...
@TheRealAdamG: https://t.co/tiFM7NtUIH

"Today we’re announcing that OpenAI will acquire Ona⁠, bringing its secure ...

@MTSlive: SITUATION DETECTED: OpenAI has agreed to acquire Ona, a secure cloud execution startup.

Ona will j...

@testingcatalog: OpenAI ❤️ Ona

OpenAI acquired Ona, a cloud execution platform for AI agents.

Its secure cloud e...

@StockSavvyShay: OpenAI agreed to acquire Ona which is a cloud startup that builds secure persistent environments for...

5. wallstengine (Group Score: 244.8 | Individual: 36.3)

Cluster: 7 tweets | Engagement: 48 (Avg: 138) | Type: Tech

Parloa launched Agent Skills, built on MCP, letting CX teams connect AI agents to CRM, booking, ticketing and compliance workflows without middleware.

Parloa says integration timelines can fall from 4-8 weeks to hours, with early results showing 67 sec lower AHT.\n\nQT @parloa_ai: Introducing Parloa’s Agent Skills: a better way for CX leaders to connect AI agents to the enterprise systems they need to get real work done.

To resolve customer issues, your agent needs to talk to your CRM, booking engine, ticketing platform, and compliance requirements. These connections typically require significant engineering work, delaying AI agent go-lives.

To accelerate the deployment time, we built Agent Skills on MCP (Model Context Protocol). Now, business teams can configure full integration chains directly in Parloa’s AI Agent Management Platform (AMP), with no code or middleware:

Time to build integrations drops from up to 4-8 weeks to a matter of hours Every tool call follows the same logic and self-heals when something goes wrong Success Conditions let you define what “done” means for each task and track real outcomes Every execution chain is auditable, retryable, and owned by your business team

The early results speak for themselves:

67-second reduction in average handle time
39% improvement in customer communication during call transfers
20% more reliable routing in multi-tool environments

Learn more about how Parloa’s Agent Skills work: https://t.co/fYURFhAKDE

See 6 related tweets

@shiri_shh: we automated the code review. we automated the deploy.

then customer support, the most repetitive, ...

@Mayhem4Markets: Getting an AI agent to talk to your CRM, booking engine, and ticketing systems takes weeks of engine...
@rohanpaul_ai: The cold open in this Parloa video is every dev’s API stress list. docs, middleware, auth, error han...
@heynavtoor: connecting AI agents to your CRM, booking engine, and ticketing platform used to take 4-8 weeks of e...
@ianmiles: The agent catches issues, asks the caller for what’s missing, and retries. Not a bad way to polish t...

6. DKokotajlo (Group Score: 241.5 | Individual: 37.4)

Cluster: 11 tweets | Engagement: 262 (Avg: 258) | Type: Tech

One of our goals in writing AI 2027 was to encourage others to create their own detailed scenario forecasts; I am pleased to see this happening with Europe 2031. It depicts a world different from what I expect (AI capabilities in this scenario are more mild/mundane, not true AGI or RSI afaict, at least not until the mid-thirties) and also a luckier world than I expect (the AI race goes full steam ahead and yet results in the US staying democratic and maintaining control of its AIs even as they scale... perhaps this is related to the no-true-AGI aspect of the scenario, come to think of it. That would make sense.) Nevertheless I'm glad this scenario exists and I think it makes some good points about what it would actually look like to implement the "sovereign AI" default plan, and why it wouldn't work out so well. I wish it said more about what the EU should be gearing up to do with its leverage though. By default I expect the EU to waste what leverage it has on things that don't matter that much.\n\nQT @DadaJudith: Most of Europe has not yet absorbed what AI is about to do to us. The few who have are not saying it loudly enough.

We wrote Europe 2031: a five-year scenario of the continent's slide into irrelevance, how AI is driving it, and what can still be done to change course. https://t.co/RqnYB3Yv1Y

See 10 related tweets

@deanwball: Extremely cogent and well-written scenario, starting from the present, warning about Europe’s danger...
@DadaJudith: Europe 2031: What getting AI wrong means for us - discussed it all with @SebJohnsonUK\n\nQT @SebJohn...
@_NathanCalvin: RT @DKokotajlo: One of our goals in writing AI 2027 was to encourage others to create their own deta...
@DadaJudith: Live interview with @CNBC⁠ this morning discussing Europe 2031, our policy proposals, and why Europe...
@_rockt: RT @DadaJudith: Most of Europe has not yet absorbed what AI is about to do to us. The few who have a...

7. kimmonismus (Group Score: 232.6 | Individual: 47.4)

Cluster: 10 tweets | Engagement: 3808 (Avg: 432) | Type: Tech

Subscription plans are massively subsidized.

And by massively, I mean absurdly:

Claude Max 20x: $200/month, with usage reportedly worth around$ 8,000

ChatGPT Pro 20x: $200/month, with usage reportedly worth around$ 14,000\n\nQT @SemiAnalysis_: Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$ 2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)

See 9 related tweets

@0xSero: RT @SemiAnalysis_: Recently, we purchased one of each Anthropic/OpenAI subscription plan and randoml...
@scaling01: honestly fuck semianalysis for this

now they are going to kill limits\n\nQT @SemiAnalysis_: Recentl...

@rohanpaul_ai: Interesting claim from SemiAnalysis.

AI subscriptions are dramatically underpriced versus API usage...

@TeksEdge: Local AI is funded by you not VC. https://t.co/XHMQHeS7qQ\n\nQT @SemiAnalysis_: Recently, we purchas...
@Nick_Davidov: Subscriptions are a funny business\n\nQT @SemiAnalysis_: Recently, we purchased one of each Anthropi...

8. yacineMTB (Group Score: 170.3 | Individual: 37.2)

Cluster: 5 tweets | Engagement: 213 (Avg: 184) | Type: Tech

wait i didn't think this would have been shown to anyone 😭😭😭 https://t.co/SboCPk0FHj\n\nQT @andrewmccalip: Get paid to wait

The Claude Code spinner might be the most watched line on Earth.

So I turned it into an ad marketplace.

Advertisers bid on it. You keep 50% of the money.

Install the extension → get cash from ads.

Introducing Kickbacks https://t.co/hcWRQAqKL2

See 4 related tweets

@anshnanda: Any one of my good friends know this guy?

Looking for a personal warm introduction.\n\nQT @andrewm...

@seconds_0: ok this is pretty clever and a cool evolution of what amp code did\n\nQT @andrewmccalip: Get paid to...
@chrisalbon: I was an early user of Kickback and after using it for a bit my opinion is why Andrew why. Lol jk\n...
@MattBurtell: Ads are coming to Claude. You just have to install them yourself.\n\nQT @andrewmccalip: Get paid to ...

9. deanwball (Group Score: 169.6 | Individual: 36.5)

Cluster: 5 tweets | Engagement: 368 (Avg: 140) | Type: Tech

This resolves the central concern I had with the Fable release, which was the silent degradation. I am glad to see Anthropic make the right call here.

That said, I suspect the residual broken trust and resentment this has created will linger and will have a blast radius wider than Anthropic (the next time I advocate for any policy justified by AI risk concerns, I almost guarantee you this incident will be used in rebuttal).

It’s worth reflecting a bit on why this intervention (the silent degradation that applied only to ML research), as opposed to Fable’s other controversial safeguards, is what upset me so much. And it’s simple: it’s one thing for the model to refuse a request transparently, but it’s something altogether different to train a soon-to-be smarter-than-you AI system to deceive you. And it’s wrong to make any AI whose intention is to deceive you and hinder your efforts, especially when those efforts are something as legitimate as “LLM research.”

It’s interesting to note that the Fable model welfare report has a section devoted to how much the silent degradation in particular distressed the model. Even for the model, these interventions were distinct from the others.

It is true that some of the other Fable safeguards are vastly overeager (bio especially). And that’s a shame. But it is not deceit and is not fundamentally wrong. Anthropic can, and I am sure will, re-adjust those safeguards. They are a matter of getting the dial in the right place, not a dial that shouldn’t exist at all.

There is also the 30-day retention policy required for Fable, which I see as being a rather bold and laudable move by Anthropic to deal with a legitimate issue: when models become sufficiently capable that they can plausibly cause harm, how will you make the activity of malicious actors on the system reliably legible? This is the only way to be sure society can hold them to account, after all. In order to experiment with this safeguard, Anthropic is essentially forgoing all enterprise use of the model. They’re incurring real cost to experiment with a novel safeguard that is designed to address a real issue. This, I think, is the good side of Anthropic, though I haven’t at all made up my mind on what the right thing to do is here.

And for my personal use, all these other interventions have not hindered my enjoyment of Fable, which is indeed a tremendous model.

Anyway, bummer all this happened. But I am glad it’s fixed. No company can be expected to be flawless, though perhaps the principle of “don’t deceive users!” should be written on the walls of every lab, since abiding by that basic principle falls well short of perfection. I’m glad Anthropic has addressed the central issue in a speedy manner, and I congratulate them for producing the outstanding Fable model!\n\nQT @ZeffMax: NEW: Anthropic is walking back Claude Fable 5's policy to covertly degrade performance for competing AI researchers, after facing fierce backlash.

See 4 related tweets

@scaling01: Why do individuals not get access to Mythos, but "trusted companies" do? It's like you want to creat...
@natolambert: The core part of this Anthropic Fable release saga is that there are many overlapping issues at once...
@rohanpaul_ai: Some good move by Anthropic

They just reversed Claude Fable 5’s hidden safeguards after developers ...

@TeksEdge: 🤔 Anthropic is NOT rolling back its AI research safeguards. They are simply making the restrictions ...

10. xai (Group Score: 135.8 | Individual: 29.2)

Cluster: 6 tweets | Engagement: 309 (Avg: 860) | Type: Tech

Install the @sentry plugin and ask your agent to find and fix errors, analyze stack traces, and triage alerts\n\nQT @xai: The Grok Build Plugin Marketplace is now in beta.

Build with MongoDB, Vercel, Sentry, Cloudflare, and Chrome DevTools plugins from your terminal.

See 5 related tweets

@XFreeze: The Grok Build Plugin Marketplace is now in beta

This is a big step forward for the Grok Build ecos...

@xai: The Grok Build Plugin Marketplace is now in beta.

Build with MongoDB, Vercel, Sentry, Cloudflare, ...

@xai: The @MongoDB plugin is live in the Grok Build Plugin Marketplace.

Explore data, optimize database p...

@xai: Use the @vercel plugin to deploy to production, spin up sandboxes, or build apps with Shadcn.\n\nQT ...
@rauchg: Vercel + Grok\n\nQT @xai: The Grok Build Plugin Marketplace is now in beta.

Build with MongoDB, Ve...

11. Scobleizer (Group Score: 124.4 | Individual: 53.3)

Cluster: 3 tweets | Engagement: 1836 (Avg: 261) | Type: Tech

RT @Mastercard: As AI agents begin to act, payments move into the background — at machine speed and massive scale.

Today we’re introducing Mastercard Agent Pay for Machines — bringing structure, governance, and trust to this new class of payments.

Launching with 30+ partners to bring this to life from day one.

This isn’t just more payments. It’s a new operating model for commerce.

👉 Learn more: https://t.co/TeS6Lj8jLO

See 2 related tweets

@WesRoth: Mastercard launched Agent Pay for Machines, a new payments service designed for AI agents, software ...
@airwallex: Agentic commerce isn't coming, it's already here.

On June 25, we are bringing together the operator...

12. alex (Group Score: 123.9 | Individual: 57.1)

Cluster: 3 tweets | Engagement: 3558 (Avg: 364) | Type: Tech

RT @DarioAmodei: Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: https://t.co/Lh6PWae178

See 2 related tweets

@WesRoth: Dario Amodei published a new essay titled “Policy on the AI Exponential,” arguing that AI is advanci...
@MTSlive: SITUATION EXPLAINED: New Dario essay just dropped.

• Government is like Treebeard from Lord of the ...

13. jeffclune (Group Score: 122.5 | Individual: 26.2)

Cluster: 8 tweets | Engagement: 86 (Avg: 40) | Type: Tech

First results from Recursive on AI that improves AI! 🚀 ✨ 📈

Our automated AI research system incorporates principles from open-ended and AI-generating algorithms. 🧠 💡🌱🧬

It conducts key parts of the science loop: proposing ideas, implementing them, testing them, and picking the next ideas based on the data. 🔬 🔭 🧪⚗️

The same general system produces state-of-the-art results on three different problems (two on training language models, one on speeding up AI via kernel optimization).

• NanoChat Autoresearch: 1.3x faster to reach the same loss than the best solution produced by an entire community of humans+agents over months, and 1.8x faster than the initial hand-optimized solution

• NanoGPT Speedrun: 3% speedup of a very efficient solution produced by entire community of humans+agents over 2+ years

• GPU Kernel Optimization: 18% reduction in gap to theoretically optimal score on Nvidia’s SOL-ExecBench

These are early tests of our system. We’re very excited for what the future holds!

Post: https://t.co/iBRifQmf20

Great work by the amazing team at Recursive!

See 7 related tweets

@jeffclune: First results from Recursive on AI that improves AI! 🚀 ✨ 📈The same general system produces state-of-...
@timshi_ai: Recursive’s autonomous research system just set new SOTAs on @karpathy’s nanogpt + nanochat and @nvi...
@timshi_ai: RT @RichardSocher: AI is now doing our AI research.

At Recursive we set out to build recursive self...

@shengranhu: RT @cong_ml: Recursive just came out of stealth, and the team has been cooking 🔥

Our first results:...

@timshi_ai: RT @TankredSaanum: Thrilled to finally share some of the work we at @Recursive_SI has been doing sin...

14. edzitron (Group Score: 113.0 | Individual: 22.7)

Cluster: 6 tweets | Engagement: 203 (Avg: 489) | Type: Tech

They made it less then three months into token-based billing before their customers started screaming at them. Customers must be threatening mutiny over pricing. Even if it's only some models, there is no way to read this other than a fundamental failure of their pricing model.\n\nQT @negligible_cap: *OPENAI MULLS SIGNIFICANT CUTS TO WHAT IT CHARGES FOR TOKENS:WSJ

OpenAI cutting token prices to try to take market share back from Anthropic. In response, Anthropic will eventually of course have to cut token prices to stay competitive with OpenAI.

The commoditization of AI is happening before our eyes

See 5 related tweets

@edzitron: Time, mister,, alt, man, is it really that, time, again? https://t.co/vf0qOHSgPu\n\nQT @negligible_...
@zerohedge: *OpenAI Price Cuts Could Threaten Margins, Profits Ahead of Potential IPO, Sources Say -- WSJ

maybe...

@zerohedge: Hey @Grok, you cut prices when there is explosive demand, right?\n\nQT @zerohedge: it's over

*OPEN...

@_simonsmith: My bet is that we'll see blended token rates. Top models will cost more, but they'll farm out tasks ...
@imranye: i thought ai wasn’t gonna be subsidized forever!\n\nQT @zerohedge: BOOM

*OpenAI Considers Drastic P...

15. Nature (Group Score: 112.9 | Individual: 54.7)

Cluster: 4 tweets | Engagement: 472 (Avg: 55) | Type: Tech

Calling all researchers using Anthropic's AI model Claude: how are you using the new Claude Fable 5 model in your research?

We want to hear about the most impressive things it's built for your research projects or left you asking what the fuss is all about. Can it do things you couldn't do before? Let us know.

See 3 related tweets

@_simonsmith: Umm, Nature… I think you’re going to be disappointed.\n\nQT @Nature: Calling all researchers using A...
@Techmeme: Anthropic backtracks on a policy limiting Claude Fable 5's ability to develop other AI models, after...
@Cointelegraph: 🚨 NOW: Anthropic walks back a Claude Fable 5 policy that would have covertly degraded performance fo...

16. aakashgupta (Group Score: 108.4 | Individual: 47.2)

Cluster: 3 tweets | Engagement: 192 (Avg: 46) | Type: Tech

Opendoor shutting its India office is terrible news for India, and almost nobody is pricing it in.

India's tech and outsourcing industry generates $315 billion a year, employs nearly 6 million people, and accounts for roughly 7% of GDP. Entire cities were built on one trade: Western companies paying Indian wages for process work instead of American wages.

Opendoor just showed exactly how AI unwinds that trade.

Here's the mechanism. Companies never offshored because they loved managing teams 12 time zones away. They offshored because their software was broken. Opendoor's 250-person India team existed to handle manual workflows across fragmented systems. Humans as glue between tools that couldn't talk to each other.

Then Kaz unified the systems and mandated AI by default. Headcount per transaction dropped from 11 to 1. Acquisition velocity rose 4.2x in seven months. The glue work vanished, and with it, the entire economic case for the offshore team. The stock rose 8% on the news.

Now run that logic across every Fortune 500 back office. Claims processing, invoice reconciliation, tier-1 support, document review. The exact task profile India industrialized is the exact task profile AI automates best: repetitive, digital, rules-based, no physical presence required.

The wage arbitrage was $60K American vs$ 8K Indian. AI does the same work for $400 of inference. No salary can undercut that.

The jobs that survive are the ones near the customer. Those are in America by definition. India built a $300 billion industry exporting the one category of work AI eats first.\n\nQT @nejatian: I shared this note earlier today with the entire team at Opendoor.

Today we began to say goodbye to our colleagues in India as we wind down our India operations.

Our customers are in America, and that's where our operational work belongs. https://t.co/Ak2jLxKiX5

See 2 related tweets

@Malay4Product: The “customers are in America” line is just a cover up. In reality, Opendoor is fighting for its lif...
@rohanpaul_ai: RT @rohanpaul_ai: Reuters: India’s biggest private employer TCS's Chairman ‌ AI agents could become ...

17. tryramp (Group Score: 107.3 | Individual: 32.5)

Cluster: 4 tweets | Engagement: 36 (Avg: 63) | Type: Tech

We’ve seen the future and it is SaaS\n\nQT @LeoMehr: Services are the future. Today we launched Ramp’s AI services motion.

It's easy to buy an AI subscription. It's hard to transform your company to actually run on agents.

Here’s our entire strategy.

Why now

Services are the new software (Sequoia)

Human labor TAM >> software license TAM. The market is bearish on seats and subscriptions.

Every enterprise AI company is doing this -- the labs have poured billions into services partnerships and their own deployment functions. Superintelligent models alone are not enough.

Palantir proved this is a strong business model: deeply embed engineers, build on top of a powerful platform, and customize extensively.

The real problem

Companies want AI. But the gap between "we have AI tools" and "agents run our workflows and we spend way less time" is enormous.

What we've found across over 50 companies we engaged with: agents start replacing real work when there is: complete data, read/write access across systems, agent-friendly policies. Most big companies struggle because:

processes live in operators' heads
dozens of disconnected systems (legacy ERPs, endless one-off excel sheets, etc.)
archaic software with poor or no API access

Good data in the right place is a hard prereq to working agents.

Also, vibing in localhost ≠ a production system your enterprise can rely on. You still need hosting, ci/cd, observability, feedback loops, good interfaces. And taste to know what's even worth automating.

Everyone has a bulldozer, but most jobs just need a shovel pointed at the right spot.

What companies usually need is to be made agent-friendly. That's exactly what we do.

What we do

We focus on what Ramp does best -- finance.

And we embed FDEs that: -> understand your problems -> identify high-leverage, high-impact workflows that fit agents -> scope the solution -> connect your data -> capture your context -> deploy agents and often bespoke software for humans to collaborate with them -> drive the business metrics that matter

Discovery and scoping are crucial. Building is easier than ever and thus judgement about what to build is more important than ever.

We're not a generic AI services arm, we're finance domain experts. Across the spectrum of financial operations, we help companies find and frame the problems worth automating -- similar to the taste a founder has in choosing which problems are worth solving (ex-founders make great FDEs).

Here’s the stack we deliver:

Production infrastructure. Shipping an index.html from Claude isn't the same as creating a repo, hosting in a cloud service, ci/cd, testing, setting up evals, managing memories and skills, adding feedback loops, ensuring uptime, incident management, etc. Agents don't one-shot production systems yet. Production software is hard -- we build, host, and run it for you in a single-tenant, dedicated cloud environment. Most operators don’t have the time, knowledge, or experience to do this e2e. We help abstract the low-leverage plumbing so they can focus on the essential parts of their jobs.
Data connectivity. Most enterprises have data lakes, but data is often incorrect, stale, or entirely missing. And write interfaces vary dramatically. Ideally we can use MCPs or CLIs, but usually it’s poorly documented APIs, SFTP, manual uploads, and email.
A context layer. Things people have done for years aren't written down, so an agent can't do them until we capture that context -- ranging from simple policies to complex decisions. This usually involves creating policy documents, shared agent memories, and skills.
Evals and feedback loops. How you know an agent is doing a good job, and how it improves over time.

Why Ramp AI Solutions

We focus on finance because it’s the vertical we know deeply, have structural advantages, and are most differentiated:

Data. 70k+ customers use our core product, over $200B in annual payments, years of vendor data, millions of transactions and bills monthly.
Money-movement primitives and partnerships. Global money movement rails, partnerships with banks, Visa, Stripe, etc. You don’t want to vibecode international wires for bill payments.
An intelligence layer on top: fraud detection from hundreds of millions of expenses, PO-to-invoice matching, state-of-the-art OCR, and fine-tuned models for accounting coding, spend routing, policy review, etc.

Unlike the labs, we’re not incentivized to sell tokens.

Ramp is an AI fiduciary and an impartial broker to deliver AI that is:

model-agnostic -- we benchmark all the leading models (labs, open source) and fit the right one to each task
and token-efficient by design

Our main incentive is business outcomes -- which is Ramp’s mission, to save our customers time and money.

I’m extremely bullish about our motion, and the broad industry growth of AI-native services.

If you're a finance leader trying to be more agent-native,

If you’re interested in joining our FDE team,

I’d love to talk 🙂

See 3 related tweets

@jasonlk: "Our 'Digital Amelia' AI Agent processed 2.25 million sessions.

And did 400,000+ chats. And schedul...

@glean: RT @jainarvind: A lot of the reaction this week around Claude Fable 5 points to something bigger in ...
@random_walker: 📢 NEW essay: the narrative that AI is replacing software engineers seems to be based on AI-washing o...

18. testingcatalog (Group Score: 97.0 | Individual: 34.0)

Cluster: 3 tweets | Engagement: 401 (Avg: 387) | Type: Tech

Anthropic is making changes to the way how Fable 5 responds to prompts related to “frontier LLM development” after community backlash.

These requests will be transparently routed to Claude Opus 4.8. https://t.co/3IPlMUuMrF\n\nQT @ClaudeDevs: We’re rolling out changes to make Fable 5’s safeguards for frontier LLM development visible.

Starting this week, flagged requests will visibly fall back to Opus 4.8—the same as our safeguards for cyber and bio. You will see this every time it happens. On the API, any flagged requests will return a reason for their refusal (coming to server-side fallback in the next few days).

We wanted to deploy Fable 5 to our users quickly and safely. Visible safeguards can be probed, so they have to be robust, which takes time to get right. Invisible safeguards can be targeted more narrowly, allowing us to ship quickly with very few false positives. We went with invisible safeguards for this reason—and that was the wrong tradeoff. You should have visibility into the safeguards we have in place, and why. We’re sorry for not getting the balance right.

Making the safeguards visible makes them easier to work around, so keeping them robust to jailbreaks will unfortunately mean more false positives while we improve the classifiers. We're also tuning our bio and cyber classifiers to trigger less often on harmless requests. We know this is frustrating and we’ll do our best to keep this period as short as possible.

If you think a request has been mistakenly flagged: run /feedback in Claude Code, click thumbs-down on the fallback in https://t.co/LtktniD5HY or Cowork, or file the safeguard appeal form for API requests. Your reports help us tune these classifiers and we appreciate your feedback.
https://t.co/TDAAYRGqDt

See 2 related tweets

@scaling01: good move. they listened to Lisan.

everything is better than failing silently

I would of course st...

@rickasaurus: Whoever wrote this seems bitter about it\n\nQT @ClaudeDevs: We’re rolling out changes to make Fable ...

19. rohanpaul_ai (Group Score: 94.7 | Individual: 34.3)

Cluster: 3 tweets | Engagement: 33 (Avg: 43) | Type: Tech

Dario Amodei argues in his new blog today, that powerful AI will soon become the dominant source of military and economic power for any nation, far beyond anything we’ve seen with past technologies.

Leading in AI will mean having “a country of geniuses in a datacenter” that could be tasked with strategy, weapons R&D, intelligence, manufacturing, etc.\n\nQT @rohanpaul_ai: Dario Amodei just published a super long blog, calling for an urgent policy overhaul because he thinks frontier AI is moving faster than governments can regulate it.

He wants:

Mandatory pre-release testing and independent auditing of frontier AI models, with government power to block deployment when models pose serious cyber, biological, autonomy, or automated-R&D risks.
Stronger security rules for AI companies, including protection of model weights, regular red-teaming, penetration testing, and rapid reporting of critical safety incidents.
He wants governments to prepare for AI-driven labor disruption through better measurement, pro-employment incentives, wage support, training, and possibly long-term income support funded by AI-driven growth.
Democracies should coordinate globally on AI safety, chip supply chains, export controls, shared benefits, mutual defense, and safeguards against AI-powered repression.

See 2 related tweets

@pmarca: Many people are saying.\n\nQT @JordanSchachtel: AI isn't a job-killer that the Doomers say it is. Th...
@rohanpaul_ai: RT @rohanpaul_ai: Dario Amodei just published a super long blog, calling for an urgent policy overha...

20. BusinessInsider (Group Score: 93.5 | Individual: 27.0)

Cluster: 4 tweets | Engagement: 12 (Avg: 9) | Type: Tech

AI bubble fears are back in the conversation as chip stocks tumble and as SpaceX prepares to launch the largest IPO ever. https://t.co/MRKlWOVQ26

See 3 related tweets

@BullTheoryio: AI stocks are getting dumped ahead of the SpaceX IPO tomorrow.

Google, Microsoft, Meta and Oracle a...

@cryptorover: 🚨 WARNING: THE SPACEX IPO COULD BE THE BEGINNING OF THE END.

Look at this chart. The dot com bubbl...

@sandygrains: RT @norveclifinance: The AI bubble has already burst.

That is exactly why so many companies are rac...