usaar333

> The only way it will result in less work and the same or better quality of life is if you’re self-employed or are employed by a compassionate human. Sadly that does not apply to the majority of people.

Even that is unlikely. If you are self-employed in any remotely competitive industry, everyone is continuing to work the same amount and AI is boosting their productivity. If you just start slacking off, your product will become inferior.

This then extends to the employment situation as well.

Only way you can get less work is if you (or your employer) are in a monopoly with a strong moat.

That's not going to stay stable forever.

> You can get 99% of jobs and work done with Haiku/Luna in a collaberating working enviroment.

Optimally? Opus will pay for itself if you save just 10% of your time

Per the charts, there is largely no point to using Sonnet 5.5 at high+ as opus low generally will give similar performance at similar or lower cost.

But Sonnet 5.5 at medium and below gives you a cheaper option at a performance worse than the lowest thinking Opus (low), which may be viable for "low intelligence" use cases.

Yeah that's not big tech. $1 trillion market cap is big tech or 30k+ employees.

This is upper end of mid sized.

No yc alum counts as big tech

Since when did doordash count as big tech?

It's sub-Fable 5 on mirror code: https://epoch.ai/benchmarks/mirrorcode?view=graph&tab=leader...

It was in one paragraph mentioned, without much detail. But yes, I agree it is one of the largest barriers.

Where's that stat from?

https://www.pewresearch.org/internet/2026/06/17/americans-an...

That's 30% positive on productive/informed, and only ~5% on hinders.

Whether you think AI will turn out bad (say by killing us all) is a different question from whether it is positive or negative at the moment.

> and eventually AI model will be commodified

This axiom not being true (and I'd bet against it) means your overall conclusion is false.

Tech workers get paid in equity and many in the semiconductor industry are making far far more than this a year with all the equity appreciation.

How does delaying the release not solve anything? It puts everyone on a notice to fix all security vulnerabilities now

I hate waiting on hold for 30 minutes even more.

There's literally a link on the blog post to an article noting they hit $150M ARR.

Voice agents have capabilities and policy to alter customer state. Just the other day I called into a CC company and the AI waived an interest charge.

page is updated to state:

MCP-Atlas: The Opus 4.6 score has been updated to reflect revised grading methodology from Scale AI.

> But even setting aside the leaked answers, the scorer’s normalize_str function strips ALL whitespace, ALL punctuation, and lowercases everything before comparison. This means:

I don't understand the concern here

True, but it gets you higher accuracy. Gemini had the best aa-omniscience score

https://artificialanalysis.ai/evaluations/omniscience

Openai has; they don't even mention score on gpt-5.3-codex.

On the other hand, it is their own verified benchmark, which is telling.

i'd interpret that as rounding error. that is unchanged

swe-bench seems really hard once you are above 80%

In Quebec it was a 20% jump in mother employment: https://www.bloomberg.com/news/articles/2018-12-31/affordabl...

And had all sorts of negative outcomes for the kids: https://www.edweek.org/teaching-learning/long-term-study-of-...

claude 4.5 gets 82% on their own highly customized scaffolding. (parallel compute with a scoring function). That beats Doubao

That wasn't a ceasefire violation. It was a six week ceasefire that had expired at the beginning of March

Physics seems better than veo 3 at least from demo videos

Except it is sublinear. Sonnet 4 was 10.2% above sonnet 3.7 after 3 months.

No it doesn't. If it were even linear compared to o1 -> o3, we'd be at 2.43 hours. Instead we're only at 2.29.

Exponential would be at 3.6 hours

No, this is below expectations on both Manifold and lesswrong (https://www.lesswrong.com/posts/FG54euEAesRkSZuJN/ryan_green...). Median was ~2.75 hours on both (which already represented a bearish slowdown).

Not massively off -- manifold yesterday implied odds this low were ~35%. 30% before Claude Opus 4.1 came out which updated expected agentic coding abilities downward.

At this point the prediction for SWE bench (85% by end of this month) is not materializing. We're actually quite far away.

No obvious gains I feel from quick chats, but too early to tell.

These benchmark gains aren't that high, so I doubt it is that obvious.

> Firstly, if your prior is that every previous startup failed, what does that say about your future chances of success?

The prior is the market. It isn't sane to use your own prior experience. (Works both ways -- if your last startup did great, shouldn't assume next will).

> 4% of YC companies become unicorns. How many startups do you need to work for before you become part of the 4%? That number is not a feasible number of jobs for one lifetime.

The bar (and what the model is calculating) is Series A from top VC, not YC Seed funding. That significantly increases odds. Specifically, ~45% YC companies get Series A, so it's more like 10% chance of a YC Series A funded company becoming a unicorn (https://www.lennysnewsletter.com/p/pulling-back-the-curtain-...).

Model is change jobs every 18 months if not booming. A 1 in 10 chance is quite reasonable over a career.

I agree there is an issue with the event being too rare, but you can't just look only at modal returns. 2/3 chance of $0 (the modal return) and 1/3 chance of $10 million profit is still pretty good odds to work with.