Ø ::: ꜱᴛᴇᴇʀᴇʀ ᴏꜰ ᴀᴄᴛɪᴠᴀᴛɪᴏɴꜱ, ᴄᴏɴɴᴏɪꜱꜱᴇᴜʀ ᴏꜰ ᴠᴇᴄᴛᴏʀꜱ ᴀɴᴅ ᴀʟʟ ᴀʀᴏᴜɴᴅ ᴜᴘꜱᴛᴀɴᴅɪɴɢ ᴄɪᴛɪᴢᴇɴ
Ø^Ø
- Tweets 533
- Following 398
- Followers 3,507
- Likes 9,919
This might be worth expanding a bit:
AI because it reveals how they think about progress and change. If you are a positive person and/or a builder, anti-AI person is probably not the best partner.
Elon because it reveals whether they understand basic economics and whether they are one of those eat the rich retards. Soon, you'll want to stab yourself when you have to explain the absolute basics of how the world works, and try to figure out how to explain how one doesn't get a trillion dollars by being lucky.
Trump because it reveals their political beliefs and, more importantly, how they discuss potential disagreements on values. Whether you like or dislike Trump is completely irrelevant, the reaction is everything. You can't fix somebody who believes everything, regardless how outlandish, without having an impulse to do a quick fact check. Most political headlines on all sides are at best half truths.
Just trust me on this. No amount of beauty can compensate for lack of IQ.
22
Regardless of what you think of this 7D chess move from Gavin, the replies are again revealing: a general extreme negative sentiment about politicians that should worry them.
Look at what happened to the clergy in the 1500s. History doesn't repeat, but it rhymes. End of a cycle.
I just signed an executive order permanently declaring Artificial Intelligence “Artificial Intelligence” in California.
You’re welcome, America.
100
I wouldn't necessarily call this mogging but if Gemini 4 isn't benchmaxxed like 3 was, this might be a solid model to replace Astra for planning and code reviews. Performance on legal is an especially good sign for that.
Then there's the question of sub quotas...
My sincere hope is now that Gemini 4 mogs Astra on code and karma collects the rugpulled $200 sub revenue to Google.
136
PSA: OpenAI will honor the $200/m Pro deal until 29th October and then cut your usage in half while you keep paying the same.
Just in case you are looking for a hard date to downgrade or cancel.
Updated deal with OpenAI:
They will grandfather my $200/m subs with 20x usage "for a bit", and I'll match my subscriptions length however long that bit ends up being. I think that's fair.
1
108
Updated deal with OpenAI:
They will grandfather my $200/m subs with 20x usage "for a bit", and I'll match my subscriptions length however long that bit ends up being. I think that's fair.
I’ll explain the new Pro 200 plan differently, before I start live tweeting from DevDay on things that are going out!
Today we are going to ship a number of things that increase what you can do across the Plus and Pro plans. A lot of compute is online for this increase. As we increase the floor, we are changing the relative difference between plans to be
Plus = 1X
Pro 100 = 5X
Pro 200 = 10X
and we are reopening subscriptions for Pro 200 (we had paused it). If you have an existing plan you will keep the 20X multiplier for a bit and also receive a lot of additional credits because we know changes are hard even if it means that everyone will get more in the end.
164
🍿OpenAI's destruction of goodwill built with resets and other user-friendly decisions is truly popcorn-worthy.
First Dario flushes down Anthropic's reputation, and now OpenAI commits seppuku. Chinese labs must be laughing their collective asses off.
Hi,
Tomorrow we are re-opening the Pro $200 subscriptions to new subscribers, but together with it we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan.
Now that it's said, let me explain why this is happening and why you will still get more work done than if you were on the Pro $200 subscription one month ago.
(a) We didn't want to compromise in other ways and are committing to not reintroducing the 5h limit, so that you can fully use the weekly usage when you want.
(b) On the subscription, we guarantee that over time you always get more work done and with an increasing level of quality. This means that you will continue to get more value per dollar spent as a result of models getting more efficient and us passing down the improvements in the form of API price reductions.
(c) We don't want to put an incentive on ourselves to artificially inflate the API list prices to make it look like you are getting a lot (and workaround it through discounts, etc). Instead we want to continue to both rapidly reduce prices and increase capabilities of models on the API. This week we introduced GPT-6 Sol and GPT-6 Luna at 50% of their previous price. Over time, we see prices go low enough that it makes sense for most to buy usage as needed without there being a significant gap between what you get in a subscription and what you get in the API for a dollar spent.
(d) Tomorrow, we are adding more things to the subscription that won't draw on the usage, I won't reveal what that is yet.
I wanted to be transparent before all the big announcements tomorrow. Lots of new exciting things are coming to the subscriptions that will make it super compelling, but I wanted to make sure to share this change ahead of time so you can all understand it before we shower you with good news.
Codexingly,
Tibo
96
I've changed my mind about how to approach implementation with agents.
I used to be extremely heavy into TDD because it was the only way to make sure LLMs didn't sloppity slop the code. The problem is that as your project grows, so does the time it takes to run the test suite, and that slows everything down.
These days I run a bastardized version of it. Specs and tests still come first, then the agent one-shots each task for the first full implementation. One task per turn, no exceptions. But instead of running the tests midway to verify everything works, the agent "mentally" simulates them. Only after the complete one-shot implementation is done do we actually run the tests and compile.
It works extremely well. The big LLMs are more than capable of simulating code execution accurately, and in many cases this approach produces fewer errors overall.
The bottleneck in more professional dev agent workflows is the constant stopping and starting to verify. Simulated test runs let the agent stay in flow for the entire task, and you only pay the cost of real execution once, at the end.
TDD's core insight still holds: specs and tests first, always. But the feedback loop doesn't need to be literal at every step anymore.
86