sangnoir

> I _strongly_ disagree with a fully cynical response of working only to contract, leveraging job offers for raises, etc

Would you consider employers to be "fully cynical" about their affairs and interactions with employees? I do. Being a happy little cog is it's own reward, but ine has to be clear-eyed about it.

> If people know you are a standout contributor then you will be hired quickly into desirable roles.

You are presenting a false dichotomy - one can be an outstanding contributor while working 40 hours per week.

> They're also hosting in switzerland.

Crypto AG was also based in Switzerland - that by itself mattered very little because it was owned by the CIA and the West Germany Intelligence service.

> Did we not watch the same movie?

"My team and I are proud to share with the world something special we've been working hard on. We are finally ready to launch the Torture Nexus, inspired by the unforgettable technology in the classic scifi novel 'Don't create the Torture Nexus'"

> And we know exactly why Meta is doing this and it’s not because they have some grand scheme to build up AI. It’s to keep these people away from their competition

I don't see how you can confidently say this when AI researchers and engineers are remunerated very well across the board and people are moving across companies all the time, if the plan is as you described it, it is clearly not working.

Zuckerberg seems confident they'll have an AI-equivalent of a mid-level engineer later this year, can you imagine how much money Meta can save by replacing a fraction of its (well-paid) engineers with fixed Capex + electric bill?

> it is trivial to spoof user-agent

It's also trivial to detect spoofed user agents via fingerprinting. The best defense against scrapers is done in layers, with user-agent name block as the bare minimum.

A curious disclaimer on the listing was that they were selling the physical plane with the Hydrogen engine, but not the intellectual property. I suppose the IP was sold separately, and the buyer might need to obtain a license separately before legally operating the aircraft.

> Is anyone actually getting value out of these models?

I've found incredible value in having LLMs help me write unit tests. The quality of the test code is far from perfect, but AI tooling - Claude Sonnet specifically - is good at coming up with reasonable unit test cases after I've written the code under test (sue me, TDD zealots). I probably have to fix 30% of the tests and expand the test cases, but I'd say it cuts the number if test code lines I author by more than 80%. This has decreased the friction so much, I've added Continuous Integration to small, years-old personal projects that had no tests before.

I've found lesser value with refactoring and adding code docs, but that's more of autocomplete++ using natural language rather than AST-derived code.

The particular startup in the article was developing tech to replace regular aircraft engines with those running on Hydrogen. I don't think any lessor would appreciate those extensive modifications on their aircraft.

Is it the media or unfettered capitalism? Traditional media (newspapers, and local news) is dying slowly, they cannot successfully compete with centralized juggernauts with global footprints sucking up all the ad dollars. Was it Thiel who remarked that capitalism may not be compatible with democracy? One guess on which he'll choose to keep.

> Sony and Microsoft don't need to interact with AMD's software, they just use the hardware

That's the entire point of AMD partnering with larger companies, rather than going all-in with consumers and small startups at this point in time.

4. The US government can chip in via grants, tax breaks or contracts.

It's all very Dr. Strangelove. "Mr. President, we must not allow an AI gap! Now give us billions"

xAI did/does the same, but Grok is nowhere near as good. Perhaps a measure of talent is required to "copy" as well as DeepSeek.

No. https://geohot.github.io//blog/jekyll/update/2022/10/29/the-...

> But Hotz isn't crazy to test them in an odd way..

He should have just read the Lisa Su interview from Q1 2024 where ahe laid out AMDs strategy without equivocating

> ... although he'd probably be better off dropping AMD cards

I think this is what's best for everyone. Looking at his recent track record[1], he seems like a person who's gets really excited by kicking things off and experiencing the exponentially growth phase, and then when it flattens out into a sigmoid curve, he dusts his hands and declares his work done, and moves to the next thing.

. 1. Hired by Elon to "fix" Twitter, CommaAI, and soon, Tiny

> That said I'm deeply worried about anyone whose based their company on amd gpus

Sony Interactive and Microsoft XBox seem to be doing great without an army of underpaid students. AMD does great at the top and bottom: the corporates in the middle that are unwilling or unable to pay people to author/tweak their software for AMD GPUs will do better going with Nvidia, which has great OOTB software, and a premium to go with it.

I suppose if AMD had infinite resources, it'd fix this post-haste.

> ...he'd be much more worried about whether AMD management have any interest in companies like his succeeding.

This reads as incredibly entitled. AMD owes him nothing, especially if he's opposed to the leadership's vision[1] and being belligerent about it.

There is maybe 1 or 2 companies with enough cachet to demand management changes at a supplier like AMD - and they have market caps in the trillions.

1. Lisa Su hasn't been shy about AMD being all about partnering with large partners who can move volume. My interpretation of this is AMD prefers dealing with Sony, Microsoft, hyperscalers, and HPC builders, then possibly tier II OEMs. Small startups are probably much further down the line, close to consumers at the tail end of AMD's attention queue. I don't like it as a consumer, but it seems like a sound strategy since the partners will shoulder most of the software effort, which is a weakness AMD has against Nvidia. They can focus on cranking out ok-to-great hardware at more-than-ok prices and build up a warchest for future investments, and who knows when this hype bubble will burst and take VC dollars with it, or someone invents an architecture that's less demanding on compute (if you're more optimistic)

> Their preference to shut down instead of receiving tens of billions of dollars would be a clear violation of a company’s fiduciary duty to shareholders for any normal company.

This is not strictly true - when a company leaves a huge market, it is imprudent to leave behind a well-resourced competitor in place. If I were a ByteDance shareholder, I'd hate if it spun off TikTok America LLC, and then having TikTok America compete against ByteDance in Europe and the Rest of the world on an equal technological footing, but perhaps even deeper pockets from American markets.

> ...two days for a traditional NLP pipeline

Why 2 days? Machine Learning took over the NLP space 10-15 years ago, so the comparison is between small, performant task-specific models versus LLMs. There is no reason to believe the "traditional" NLP pipelines are inherently slower than Large Language Models, and they aren't.

> Yet people are arguing that we should allow the people who did that to continue to install apps on millions of Americans phones

Who are "the people who did that" - Byte Dance or China as a whole? If it's the latter, I'm afraid there are still plenty of apps made by Chinese companies like, DJI, Lenovo, and thousands of IoT apps to control random geegaws via WiFi or BT.

It's not hard to see the pattern: any Chinese tech champion that does as well as, or better than American companies will find itself in legal peril. Huawei didn't get in trouble after hacking Nortel, but they got sanctioned much later, when their 5G base equipment was well-received by the markets. TikTok had the best ML-based recommendation systems when it burst in the scene, Google and Meta still haven't quite caught up yet.

> Yet people are arguing that we should allow the people who did that to continue to install apps on millions of Americans phones

Who are "the people who did that" - Byte Dance or China as a whole? If it's the latter, I'm afraid there are still plenty of apps made by Chinese companies like, DJI, Lenovo, and thousands of IoT apps to control random geegaws via WiFi or BT.

It's a reference to a Monty Python sketch - What Have the Romans Ever Done For Us? https://youtu.be/Qc7HmhrgTuQ?si=M_0VYwJBXwkzL1qV

What Has Google Ever Done for AI Research?

The comment at the root of this thread was specifically addressing making it to orbit on first launch - which a bench test isn't.

TBF: that wasn't an unsuccessful launch attempt, but a failure to not launch. Which affirms parent in that they seem to have work out all the kinks out during development.

Would you ask the same question if it were Long March rocket debris falling into the Atlantic with very short notices from China?

> ...in the early days it was just a hodgepodge mess of ideas, even if it dominated the market.

and it was glorious; the intent-system and Notifications drawers were Androids calling card. Intents were a blessing and a curse: being able to replace apps was great, but the variety in design language, not so much.

Being able to reach into apps' storage was insecure, but freeing one's data from SQLite files was fantastic.

How about nouns as verbs? "The new dashboard will surface potential issues. If we find any ,I will calendar a meeting for the cross-functional group to workshop the list, and task the relevant partner-teams to resolve"

I think the American government is contorting its public argument to avoid saying this plainly because there are many American companies that control information for most of the world, and they don't want other countries to go "Hmm, hang on a minute..."

> the abundance of blocking in the 'verse in lieu of adequate spam controls, which ActivityPub fundamentally lacks

Bring your own algorithm can tackle spam (like adblocks). ActivityPub is flexible enough for Mastodon to build up references to algo-providers.

I've been frustrated by Mastodons slow movement on this, I considered approaching the popular clients to implement this. After all, ad-blockers run entirely on the client. A basic standard can be drawn up that governs how to boost or penalize toots' visibility based on keywords, author or instance, based on an updated list is viable today amd can run entirely on the user's device. Web-based clients are more complex, and require patching the standard Masto server or running in a sidecar.

> the abundance of blocking in the 'verse in lieu of adequate spam controls, which ActivityPub fundamentally lacks

Bring your own algorithm can tackle spam (like adblocks). ActivityPub is flexible enough for Mastodon to build up references to algo-providers.

I've been frustrated by Mastodons slow movement on this, I considered approaching the popular clients to implement this. After all, ad-blockers run entirely on the client. A basic standard to boost or penalize toots' visibility based on keywords, author or instance based on an updated list is viable today.