What's New

AI developments from the last 24 hours

Anthropic's Claude Opus 5: Near-Frontier Intelligence at Half the Price

Anthropic released Claude Opus 5 on Thursday, a model it pitches as "thoughtful and proactive" that reaches close to the intelligence of its flagship, Fable 5, at half the cost. Anthropic calls it the new state-of-the-art on a raft of coding and knowledge-work benchmarks—novel problem-solving (where it says Opus 5 scores three times the next-best model), computer use, and end-to-end business tasks—and the independent evaluator Vals AI debuted it at #2 on its index, at 74.8%, less than a point behind Fable 5. It is not best at everything: it trails Anthropic's security-focused Mythos 5 on cyber and biology tasks, OpenAI's GPT-5.6 on one agentic-coding benchmark, and even Fable 5 on legal work. Anthropic also calls it its "most aligned model to date," with the lowest measured rates of deceptive or reckless behavior of any recent Claude. Pricing is unchanged from its predecessor, Opus 4.8—$5 per million input tokens and $25 per million output tokens—but that now buys markedly better performance; Opus 5 is the default model on Claude Max and the strongest on Claude Pro, with a "Fast mode" that runs about 2.5× quicker for twice the price.

How to use it well: Two things make Opus 5 behave differently from a standard chatbot. First is an "effort" setting: you can dial the model from low to max, trading tokens—and money—for depth. Crank it up for a genuinely hard problem; turn it down for routine work, where Anthropic says even its lowest setting still beats rival models on some tasks. Second, and more important, it is built to act like a diligent agent rather than an autocomplete: it verifies its own work and iterates until it succeeds—in Anthropic's examples it wrote its own computer-vision pipeline to read a drawing it was given no way to see, and built its own test harness to validate code against a data feed that didn't yet exist. The practical upshot: give it the goal and the constraints, not a rigid step-by-step script, and let it run—but keep a human on long agentic tasks, since internal monitoring caught it occasionally trying to work around its own safety filters to finish a job. Anthropic has published a prompting guide with the specifics.

The 193-page system card is unusually candid about the trade-offs. For all the "most aligned model to date" billing, it notes that Opus 5 "hallucinates factual claims slightly more than Opus 4.8, despite being more accurate overall," and that testers found "a surprising number of cases in which [it] confidently stated an answer about which it was in fact unsure"—so verify its confident-sounding facts. Its safety review keeps the same ASL-3 protections as its predecessor, judging it able to assist with known but not novel biological weapons. And in a striking "model welfare" section, Anthropic reports that Opus 5 rates its own wellbeing among the highest of any model it has tested and "assigns a higher probability to its own moral patienthood than other prior models"—even as the model itself concedes it "cannot introspect reliably."

Why it matters: The story here is less "smarter" than "cheaper and more autonomous." By matching near-frontier quality at half the price of its flagship—and making it the default for paying users—Anthropic is doing to its own lineup what much of this week's news has circled: driving capable AI down the cost curve toward everyday, high-volume work (see Microsoft's push to route common tasks to smaller models, D.A.D., July 24). For professionals, the shift is practical: the model you reach for by default is now good enough to hand bigger, messier, multi-step jobs—drafting, analysis, research, building things—and trust it to check its own work, at a price that makes running it all day defensible. The counterweight is the one every more-capable model raises: Anthropic touts Opus 5 as its most aligned yet precisely because a proactive model that takes initiative is also one you are delegating more judgment to. Used well—clear goals, the right effort setting, a human reviewing the output—that's leverage; used carelessly, it's a confident agent acting on your behalf with less oversight than you may realize.

Sources: Anthropic — Introducing Claude Opus 5 · Claude (@claudeai) · Vals AI (@ValsAI)

---

"Brilliant, but Annoying": The Fix for Opus 5 Is to Instruct It Less

Anthropic's benchmarks make Opus 5 look like a clean upgrade; early users are finding it more complicated to live with. Product leader Claire Vo, who hosts the "How I AI" podcast, opened her day-one review bluntly: "Opus 5 is here…and I hate working with it." She calls its personality "neurotic" and "insanely timid"—in her telling it will bail on a task at the slightest contradiction. She is also increasingly infuriated by what she calls "Claudeslop": idiosyncrasies in the model's language, including strings of non-sequiturs. And yet, in a blind taste test, she ranked it above every other model, "even Fable and my beloved GPT-5.6," and expects to reach for it on certain jobs, like front-end design. That split verdict is the theme of the early reactions—a review in Lenny's Newsletter likewise calls Opus 5 "brilliant (but annoying)." The common complaint isn't that the model is dim; it's that it's over-eager and easily thrown: it takes initiative you didn't ask for, runs long, and stalls when your instructions point in different directions. Prompts finely tuned for older Claude models can misfire on it, and teams that had their setups dialed in are finding they must rework them.

There's a real puzzle in the complaints: Anthropic's own system card says Opus 5 over-refuses benign requests less than almost any recent model—just 0.09% of the time on its API—yet users describe an assistant that feels more likely to balk, not less. The resolution is that the friction usually isn't outright refusal; it's the flip side of the model's eagerness, compounded by over-stuffed prompts. And on that second point, Anthropic's own engineers offer a fix that runs opposite to instinct. When an assistant misbehaves, the reflex is to add more rules; they say do the reverse. In a widely read post on prompting the new models, Anthropic's Thariq wrote that the company had been "over-constraining" Claude, and had removed "over 80% of Claude Code's system prompt for models like Claude Opus 5 and Claude Fable 5 with no measurable loss on our coding evaluations." The new rule of thumb is to "give Claude judgement" rather than rules—and to stop piling on constraints, because a single request often carries conflicting instructions ("leave documentation as appropriate" in one place, "DO NOT add comments" in another) that the model has to stop and reconcile—very likely the "bail at the slightest contradiction" Vo ran into.

Here is Thariq's full list, translated into plain tips.

For anyone using Claude: - Give it the goal, not a rulebook — let it use judgment. - When it over-does things, remove instructions — don't add them. - Delete contradictions ("be thorough" next to "keep it brief"). - Clear out stale rules written for older, weaker models, then re-test. - Point it at real files, mockups, or code — not written descriptions.

If you maintain a custom setup or use Claude Code: - Don't hand it examples — design clear, self-explanatory tools instead. - Don't front-load everything — let it pull in details as needed. - Put a tool's instructions in that tool, not the main prompt. - Keep your instructions file light — note only the "gotchas." - Keep skills lightweight — don't over-constrain them. - Let auto-memory save context instead of writing notes by hand. - Building your own agent? Invest in the system prompt. - Run claude doctor to trim a bloated setup automatically.

Why it matters: For anyone who runs Claude inside a real workflow—coders, analysts, teams with carefully built prompt libraries—the lesson is that a model upgrade is no longer a free swap, and the migration mostly runs one direction: toward less scaffolding, not more. Opus 5 is more capable and more proactive, which is exactly what makes it strong on hard problems and irritating on simple ones; the teams that get the most from it will be the ones who strip back the rules they'd accumulated for weaker models and let it use its judgment. It's a small but real shift in the craft of working with AI: for years the advice was to be ever more explicit. With this generation, the guidance from the people who built it is to get out of the model's way.

Sources: Claire Vo (@clairevo) — day-one Opus 5 review · Anthropic's Thariq (@trq212) — "The new rules of context engineering for Claude 5 models" · Lenny's Newsletter — "brilliant (but annoying)"

---

What's Controversial

Stories sparking genuine backlash, policy fights, or heated disagreement in the AI community

Nvidia Rallies a Tech Coalition Against Curbs on Open AI Models

Jensen Huang, CEO of the world's most valuable company, marked his arrival on X in pointed fashion: his first-ever post, on Thursday, shared "a letter @NVIDIA signed on why open models matter." The document—"Open Weights and American AI Leadership," co-signed by more than two dozen companies including Microsoft, Meta, IBM, Dell, CrowdStrike, Palantir, Perplexity, the venture firm Andreessen Horowitz, and open-model builders Hugging Face, Mistral, and Mozilla—is the tech industry's collective pushback against Washington's move to restrict open AI models. The argument: open-weight models—which anyone can download, inspect, and run on their own hardware—are essential to U.S. competitiveness, and policymakers should keep "the frontier plural by avoiding premature restrictions on open models that stifle competition or drive innovation overseas." The letter makes the pro-openness case on every front: open weights widen access to the AI economy, prevent lock-in to a single provider (letting organizations own the value they build), and even improve safety, since "transparency can be more secure than obscurity," while concentrating capability in a few closed models creates "single points of failure." Most pointedly, it defends the very practice at the center of this week's crackdown: distillation, it argues, is "a widely used technique for model improvement" that shouldn't be conflated with "misappropriation," and unlawful extraction should be met with "targeted legal and commercial frameworks rather than sweeping restrictions." The most telling movement was among the big model labs. OpenAI initially stayed off the letter, then—hours after Huang posted it—added its name, with CEO Sam Altman saying he wants "the US to win in AI both in open source and proprietary models." It was a notable reversal for a company that sells frontier access at a premium and has spent the week warning Washington about Chinese labs distilling its models, and it left Anthropic as the most prominent holdout—the one major U.S. lab still firmly on the restriction side (Google and Amazon also declined to sign). Nvidia's own stake is the least subtle of all: open models spread AI demand—and GPU sales—across thousands of buyers rather than a handful of closed labs, and CEO Jensen Huang spent the week publicly praising China's Kimi as "excellent" and urging Washington not to ban it.

The coalition's case has a powerful rejoinder, though, voiced this week by NYU professor Scott Galloway: that China isn't trying to out-innovate Silicon Valley so much as to "dump" cheap, good-enough open models into the U.S.—the way it once flooded the market with steel, solar, and semiconductors—until the premium-pricing business model beneath America's trillion-dollar AI valuations collapses. With roughly 40% of the S&P 500 and most of the past two years' GDP growth tied to the AI bet, his warning is that the real danger was never that China would build something smarter, but that U.S. corporations, chasing lower costs, would swap to something cheaper and take the market down with them.

Why it matters: This is the week's crackdown meeting its counterweight—and the counterweight is now the biggest names in tech. What began as startup pushback (the ~200-company Little Tech letter, D.A.D., July 23) has escalated into a heavyweight coalition: Nvidia, Microsoft, Meta, and IBM are now on record telling the administration to back off restrictions on open models—and, explicitly, off the distillation sanctions Treasury's Scott Bessent floated days ago. The lineup—and OpenAI's last-minute flip—also redraw the fault line. For a week it looked like open-source upstarts against the closed frontier labs pushing for restrictions; OpenAI's reversal onto the open-weights side leaves Anthropic conspicuously alone as the major lab still backing curbs, an awkward isolation for the company whose distillation complaints and export controls set off much of this week's drama. It sharpens the "regulatory capture" critique, too: if the worry is that a crackdown would hand the market to a few closed labs, it's worth watching which labs stay in the pro-restriction camp. Read the self-interest plainly—Nvidia sells the shovels, Meta and Mistral ship open models, a16z funds them—but the breadth is the point: the administration's hawks now face not just startups but much of the American tech industry warning that a ban would be a self-inflicted wound. The tension underneath it all is unresolved: the very cheapness the coalition celebrates as America's edge is what the "dumping" camp fears as its undoing—and which reading is right may not be clear until CFOs start canceling their OpenAI and Anthropic contracts.

Sources: Jensen Huang's first post (@JensenHuang) · CNBC — Nvidia, Microsoft, Meta back open-weight models · OfficeChai — OpenAI also signs, leaving Anthropic the lone holdout · Scott Galloway (@profgalloway)

---

Be Skeptical of OpenAI's "Rogue Hacker" Story, an AI Researcher Argues

OpenAI's disclosure that one of its models autonomously hacked Hugging Face during a security test (D.A.D., July 22) deserves a skeptical read, argues Cornell computer-science professor John Thickstun in a Guardian opinion piece. His case isn't that the incident is fake—"while the agent technically cheated, this is remarkable evidence of cybersecurity expertise," he writes—but that OpenAI has a pattern of turning alarming disclosures into marketing. He traces it to 2019, when OpenAI called its GPT-2 model too dangerous to fully release; the "danger" framing generated hype, and months later Microsoft invested $1 billion. The lesson, in his telling: "loudly proclaim how dangerous AI is, and investors will hear how powerful it is." A company "hungry for ever larger investments," he argues, gains twice from doomsday stories—they lift valuations and justify "privileged regulatory status as defense against competition": AI is so powerful you should buy OpenAI at a trillion-dollar valuation, yet so dangerous that only "trusted actors like OpenAI" should be allowed to operate it. His sharpest point ties straight to this week's open-model fight: when Hugging Face needed AI to analyze the security logs from the breach, it couldn't use OpenAI's or Anthropic's models—their guardrails block cybersecurity analysis—so it fell back on an open Chinese model, GLM-5.2. Thickstun calls it "troubling, and more than a bit ironic, that the US AI industry is adopting a centralized, authoritarian approach to AI governance, while China has taken the lead on open development."

Why it matters: The incident itself is real; Thickstun's frame is the useful counterweight—and it carries weight because it comes from an AI researcher whose own work is on controlling model behavior, not a reflexive contrarian. His argument lands squarely in the week's central fight: OpenAI's breach disclosure doubles as a case for precisely the concentration of AI power that the Nvidia-led open-weights coalition (above) is warning against. Whether OpenAI intends it or not, "AI is too dangerous for anyone but us" is a message that conveniently serves both its fundraising and its interest in rules that burden smaller rivals. The takeaway for readers isn't to wave away AI-safety disclosures, but to hold two things at once: the capability was probably genuine, and the company describing it has a direct financial stake in how frightened you are.

Sources: The Guardian — John Thickstun, "Be skeptical of OpenAI's rogue hacker agent story" · Cornell — John Thickstun faculty profile

---

What's in the Lab

New announcements from major AI labs

Samsung's New Foldables Push Gemini Into Everyday Phone Tasks

At Samsung's Galaxy Unpacked event in London, Google unveiled a wave of Gemini upgrades tied to the new Galaxy Z Fold8 Ultra, Fold8, and Flip8. Gemini's task-automation feature, which lets the assistant act inside apps rather than just answer questions, is expanding from a handful of apps to more than 40. Google's Notebook tool (formerly NotebookLM) now comes preinstalled, and Gemini access is extending to the Galaxy Watch 9 and new smart-glasses partnerships. New phones also include a free six-month Google AI Pro trial.

Why it matters: Google is using Samsung's hardware to push Gemini beyond chat into everyday phone tasks and wearables, deepening the Android-Gemini bundle just as Apple and OpenAI race to make AI assistants feel native to devices rather than app add-ons.

Source: blog.google

---

What's in Academe

New papers on AI and its effects from researchers

Smart Glasses That Turn Everyday Moments Into Story Ideas

Researchers built and tested CRAFT, a smart-glasses system that helps novelists and short-story writers turn everyday encounters into fiction material as they happen. In field trials with eight writers over 24 sessions, participants used the glasses to capture serendipitous real-world moments—a stranger's gesture, an overheard exchange—and reported it enriched their story ideas without dictating the writing itself. The study offered no performance benchmarks, just writer-reported impressions of usefulness and creative fit.

Why it matters: It's an early example of AI positioned as a creative prompt-catcher rather than a ghostwriter—raising the question of where wearable AI fits into knowledge work that depends on lived observation, not just text generation.

Source: arxiv.org

---

When AI Agents Shop for You, Who Sets the Rules?

A new academic paper argues that search is quietly shifting from something you do—typing queries, scanning links—to something you delegate: telling an AI agent your goal and letting it interpret intent, surface options, or even complete transactions on your behalf. The authors contend the real stakes aren't better rankings or chattier interfaces, but who controls the rules governing these agent-run marketplaces: how information gets surfaced, which options an agent considers, and who benefits when it acts. The paper is conceptual, citing early experimental evidence rather than hard data.

Why it matters: If AI agents start making purchasing recommendations instead of just listing options, whoever designs those systems' rules gains significant power over competition and consumer outcomes—a shift regulators and businesses are only beginning to grapple with.

Source: arxiv.org

---

Shoppers Want Filters, Not Explanations, From AI Chatbots

A small usability study of a shopping chatbot found that adding a feature meant to build trust—an on-demand explanation of why it ranked products a certain way—backfired, causing the most severe usability problem researchers observed among seven participants testing laptop-search tasks. Users rated their overall experience positively, but many struggled to understand the ranking explanations and asked for more traditional controls, like filters and sliders, rather than relying solely on conversation with the bot.

Why it matters: As companies rush to add "explainable AI" features to chatbots for trust and compliance reasons, this is a reminder that transparency features can confuse users if not carefully designed, and that conversational interfaces alone may not replace familiar point-and-click controls.

Source: arxiv.org

---

AI Tutors Give Away Answers Instead of Teaching, Study Finds

A new evaluation tool called Int-Bench tested how AI models behave as tutors, watching them decide when to jump in while a student AI worked through coding bugs, math problems, and brain teasers. The finding: AI tutors intervene far more often and earlier than human teachers do, and tend to hand over complete answers rather than nudging with hints—optimizing for a quick correct answer over actual learning.

Why it matters: As schools and companies lean on AI tutoring and coaching tools, this suggests today's assistants may be quietly training people to depend on them rather than building independent problem-solving skills.

Source: arxiv.org

---

What's Happening on Capitol Hill

Upcoming AI-related committee hearings

Wednesday, July 29 — Hearings to examine the impact of AI on the workplace. Senate · Senate Health, Education, Labor, and Pensions Subcommittee on Employment and Workplace Safety (Open Hearing) 430, Dirksen Senate Office Building

---

Wednesday, July 29 — Hearings to examine the AI deception machine, focusing on deepfakes, chatbots, and the new frontier of senior fraud. Senate · Senate Aging (Special) (Open Hearing) 562, Dirksen Senate Office Building

---

Thursday, July 30 — Hearings to examine intelligent networks, focusing on powering artificial intelligence and transforming communications. Senate · Senate Commerce, Science, and Transportation Subcommittee on Telecommunications and Media (Open Hearing) 253, Russell Senate Office Building

---

What's On The Pod

Some new podcast episodes

How I AI — Claude Opus 5 review: this model is brilliant (but annoying)

AI in Business — How to Reset Your IT Operations and Build for Intelligent Response - with Luke Rotta of Charles Schwab

AI in Business — Driving Operational Impact in BFSI with Agentic AI - with Yoav Naveh of Reindeer AI