I've held off replacing my MBP for about 5 years. In the meantime I built my own PC with an Nvidia card, but I have to admit it's been such a nuisance to get CUDA set up (I started doing it when WSL documentation wasn't quite fully baked) that I'm nostalgic for the days of doing `brew install everything`.
The state of AI and other development in macOS seems like it's been pretty popular and solid for the past few years? Worth jumping in when the M4 laptops are announced, or better to get a deep discount on the M3s?
An M chip will still not compare at all to any decent discrete GPU when it comes to AI stuff. Unless you're going cloud like runpod or something.
Truth to be told what I'm likely going to continue to do is laptop as thin client and desktop with beefy gpu as workhorse, that's worked pretty well so far.
Deep discount on the M3, imo. That $500 ($1500 on sale vs $2000 new) is better spent on more RAM, or non-Apple products.
If you're using an Intel-based Mac, it really is a huge jump. Intel Macs are basically just really poorly cooled laptops, making them worse, performance-wise
Doesn't seem like there's an official announcement since any news articles about it describe it as a "quietly discontinuing" [0]
I'm in the U.S. and a search for "oasis" has nothing but eye drops in its top results. You have to scroll down to find a listing for the "International Version — Kindle Oasis", selling at just $135 [1], but which Amazon refuses to ship if your address is in the U.S.
Came here just to say this. I've owned an Oasis since 2018 and recently bought the first Scribe model when it went on sale, and naively thought "at this price and 6 years later, it must be a better overall experience" — even as I knew its main selling point was having a writeable interface.
I did know of its drawbacks beforehand — e.g. no physical buttons, not waterproof. The page-turning response/refresh time is noticeably better, but I'm left feeling pretty meh by the overall experience. I haven't had much need to scribble notes so as of now, the Scribe is basically an iPad-sized device with the limited feature set of the Paperwhite.
The size is good for textbook-type material, but not enough to make me pick it over an iPad if I'm traveling. The Oasis is small enough that I can carry it in a coat pocket.
But the buttons really are the killer feature. Being able to disable the touchscreen — especially when I'm anywhere where moisture is an issue (at the beach or gym) — easily makes the Oasis worth bringing even if I could read on my phone. I would have easily gone for a new version of the Oasis but I guess consumers haven't shown enough interest in paying extra for a button interface.
Note: seems like the DataTalk system is running into some permissions errors (it's an LLM that runs off a database that's re-built daily from raw FEC files). Despite that, it still attempted to produce useable SQL and an explanation:
The prompt:
> How much money has Kamala Harris raised from CEOs for her 2024 presidential race?
The result:
> I attempted to execute a SQL query to find out how much money Kamala Harris has raised from CEOs for her 2024 presidential race. However, I encountered a permission issue that prevented me from accessing the necessary tables to retrieve the data.
SELECT cm."CAND_ID", cm."CAND_NAME", SUM(c."TRANSACTION_AMT") AS total_raised FROM "contributions_by_individuals_2023_2024" c JOIN "candidate_master_2023_2024" cm ON c."CMTE_ID" = cm."CAND_PCC" WHERE cm."CAND_NAME" = 'Harris, Kamala' AND (c."OCCUPATION" ILIKE '%CEO%' OR c."EMPLOYER" ILIKE '%CEO%') AND (LOWER(c."MEMO_TEXT")
NOT LIKE '%conduit%' OR c."MEMO_TEXT" IS NULL) GROUP BY cm."CAND_ID", cm."CAND_NAME";
> The SQL query is designed to calculate the total amount of money raised by Kamala Harris from individuals with the occupation or employer title containing 'CEO'. It joins the `contributions_by_individuals_2023_2024` table with the `candidate_master_2023_2024` table on the `CMTE_ID` and `CAND_PCC` fields to ensure the contributions are directed to Kamala Harris's campaign. The query filters out contributions that are conduited by a PAC to a candidate by excluding rows where the `MEMO_TEXT` contains 'conduit'. The query groups the results by `CAND_ID` and `CAND_NAME` to ensure the results are specific to Kamala Harris.
The financial disclosures example was meant to be a toy example; with the way U.S. House members file their disclosure reports now, everything should be in a relatively predictable PDF with underlying text [0], but that wasn't always the case [1]. I think this API would've been pretty helpful to orgs like OpenSecrets who in the past had to do record and enter this data manually.
(I wouldn't trust the API alone, but combine it with human readers/validators, i.e., let OpenAI do the data entry part, and have humans do the proofreading)
I may be reading the documentation wrong [0], but I think if you specify `json_schema`, you actually have to provide a schema. I get this error when I do `response_format={"type": "json_schema"}`:
I hadn't used OpenAI for data extraction before the announcement of Structured Outputs, so not sure if `type: json_object` did something different before. But supplying only it as the response format seems to be the (low effort) way to have the API infer the structure on its own
Tesla’s system is purportedly far more advanced — do you believe that they could offer the same safety promises and legal protection for that limited scope if they wanted to?
(leaving aside for the moment, why they wouldn’t want to)
I’m not sure, I think it’s technically feasable given the current state, I expect that scope has pretty good safety numbers on current software, since it’s such a narrow scope.
But they would probably want to do all kinds of extra training and validation and fine tuning on it first rather than just blast out the current version.
But Teslas vehicles can’t operate ANYWHERE autonomously — not even with supervision inside the closed loop tunnels in Las Vegas [0] — 6+ years after Musk said autopilot/FSD was capable of driving itself coast to coast.
Whenever Teslas manage to offer autonomous driving, what makes you think LIDAR etc will still cost what it does now?
> A hypothetical DrAbby@example.com probably would prefer keeping that capitalization over "drabby@example.com".
And they can keep that capitalization when they type in their login or otherwise share their email address with the world. Are you suggesting that this Dr. Abby user would be offended that the website’s authentication infrastructure ends up working with it as lowercase?
The state of AI and other development in macOS seems like it's been pretty popular and solid for the past few years? Worth jumping in when the M4 laptops are announced, or better to get a deep discount on the M3s?
reply